[HerbalistReview archives]
[RMHI Home]
[HerbalThinkTCM software] [Tutorials]
[RMHInet] [Courses/Certification] [FAQ]
[Articles] [About]
[Follow] [Contact]

— updated 2004-11-01


Herbalist Review, Issue 2004-#3:
Herb-herb and herb-drug interactions: modes of interaction

by Roger W. Wicke, Ph.D.

"Herb-drug interaction" is a term bandied about by both the medical profession and the media to induce fear and uncertainty. Here we explore this term from a rigorous mathematical and pharmacological perspective to reveal that this term is frequently misused.

Subtopics on this page…

Copyright ©2004 by RMH-Publications Trust; all rights reserved.



Herb-drug interactions are a hot topic of debate, and herbs are coming under increasing attack for being potentially dangerous to patients who are already taking prescription medications. The concerns are multiplied for those patients currently taking multiple medications, often prescribed by multiple physicians who may or may not be in communication with each other regarding their medical reasoning. Non-error, adverse effects of medications account for the fourth leading cause of death in the United States (106,000 deaths/year) •[a1]• •[a2]•, while annual deaths from herbs until 2002 were only a handful nationwide. •[b1]• •[b2]• (In the latter case, most of these deaths classified as being from "herbs" were actually from accidental ingestion of poisonous house plants.) Now that an increasing percentage of the public is taking herbs, often without their physician's knowledge, herbs have become a popular scapegoat for the high risks inherent in the allopathic medical system. Consequently, it is important for herbalists to understand the nature of herb-drug interactions, and to defend their profession and their products from unjustified attack.

A basic problem is that the phrase "herb-drug interaction" routinely appears in the media, without definition and an assumption that everyone knows what is meant. There are several distinct categories of interactions defined in pharmacology texts. The basic concept of interaction is that consumption of two substances (herbs, drugs, foods, liquids) at the same time, or within a short time of each other, will result in physiological effects that are different from a simple summation of the effects of either taken separately, which is the basic definition of a non-linear effect. The vast majority of so-called interactions are probably not interactions at all and are simple linear effects, but result from individual users' ignorance of the total known clinical effects of a substance, published in standard clinical materia medica. Another, relatively less common type of interaction is a pharmaceutical interaction in which the two substances interact by direct chemical action with each other, either before ingestion, or while mixed together in the stomach and intestines.

As an example of a pseudo-interaction, consider the example of a man who is currently on a medical prescription for heparin, an anti-coagulant, and who also decides to take the TCM formula Rhubarb and Moutan Decoction (da huang mu dan tang) for Interior-Heat type constipation. If he remains unaware of the secondary effects of this formula on invigorating Blood (three of the ingredients, Rz Rhei, S Persicae, and Cx Moutan Radicis all act as Blood-invigorators), he might be surprised by a sudden onset of bleeding disorders. FDA officials and medical authorities might also be quick to label this incident as an example of an adverse herb-drug interaction, when it is likely a simple case of linearly additive effects resulting from overdose on substances whose combined total Blood-invigorating, anti-coagulant effect is too much. A variation of this type of incident arises from the American propensity to assume that if a little is good, a lot more is better. This type of attitude may easily result in adverse herb-drug incidents that get reported as interactions, yet really are simple cases of overdosing.

In contrast to heparin, the use of coumadin as a blood thinner has been observed to be one of the most susceptible drugs to not only herb interactions, but to daily diet, which is why clotting time and platelet counts are checked so frequently in people taking this drug. Almost anything that affects liver metabolism, such as Liver Qi-regulating herbs, alcohol, NSAIDS, and many other drugs may affect the breakdown and excretion rate of circulating coumadin, thus affecting the clotting time. The case of coumadin provides examples of true drug-herb and drug-food interactions, and is often the example given to illustrate the possibility and the dangers of such interactions. However, coumadin is inherently a dangerous drug; that ordinary foods, beverages, and herbs may alter its metabolism is not evidence that these foods, beverages, and herbs are inherently dangerous.

True interactions, as exemplified by the case of coumadin, are probably rare. From my own experiences and from interacting with other clinical health professionals, the vast majority of adverse herb incidents I have witnessed result from an ignorance of the total clinical effects of an herb as summarized by the TCM clinical functions and contraindications, available in standard textbooks of TCM herbology. •[b3]• The most common example of this mistake is the blatant attempt to recommend or choose herbs based on simplistic biomedical indications, such as using the formula Rhubarb and Moutan Decoction for constipation, regardless of the associated TCM syndrome type. These types of errors are responsible for what many medical professionals and TCM practitioners carelessly refer to as herb-drug interactions. One article appearing in the American Journal of Chinese Medicine admitted, "The interactions between Chinese herbs and modern drugs are a common issue, yet reports on them are very rare." •[b4]• This same article then proceeded to outline a number of assumed and reported interactions, again failing to distinguish between true interaction and simple overdose.


Mathematical models of the different types of substance-substance interactions

To explicitly pinpoint the defects in arguments regarding herb-drug interactions, it will be necessary to define what we mean by "interaction" mathematically, and to differentiate the various types of interactions by such means. Anything short of this risks allowing confusion and poorly defined terminology to gain the upper hand in public policy debates.

To describe physiological interactions between substances (herbs and/or drugs), we first need to have a mathematical model of the effects of a single substance. The background theory for this has been outlined elsewhere •[c1]• •[c2]• and is summarized here. Briefly, to be consistent both with traditional Chinese models of herbal actions and with pharmacological models, we describe the action of an herb by a vector consisting of multiple independent (statistical and functional independence) and mutually orthogonal parameters that reflect the relative strength of action in various domains.

In the equation below, hi is the vector variable representing the net effects of an ingredient, where i is the index value indicating a specific ingredient. P = the total number of simple real-valued parameters, or variables, used to characterize the general TCM properties frequently relevant in evaluating contraindications.

(Eqn. 1)     Eqn. 1

For example, hi1 might correspond to the relative intensity of action of an herb in tonifying Qi; the remaining vector components, hi2 to hiP, would specify other properties such as ability to tonify Yin, Yang, and Blood, clear Heat, resolve Blood Stasis, resolve Dampness, and to resolve various components of Exterior disorders. For purposes of evaluating contraindications, the properties may be summarized by the herb's ability to resolve or counteract each of the following set of 16 aspects (P=16):

  • Exterior-
    • Wind
    • Deficiency
    • Excess
    • Cold
    • Heat
    • Damp
    • Dryness
  • Interior-Deficiency:
    • Deficiency of Yang
    • Deficiency of Qi
    • Deficiency of Blood
    • Deficiency of Yin
  • Interior-Excess:
    • Stagnation of Qi
    • Stagnation of Blood
    • Dampness
    • Cold
    • Heat

If an herb tends to aggravate one of the components above, the value corresponding to that component would be negative. (The principle of using vectors to describe sets of variables which may or may not interact is just as applicable to sets of objectively obtained physiological variables that are measured in lab tests.)


Linearly additive, no interaction

Using the preceding foundation to describe the clinical situation in which N medicinal ingredients (herbs and/or drugs) are combined but do not interact, and whose effects sum linearly:

(Eqn. 2)     Eqn. 2

where the qi are weighting factors depending upon the amounts of each herb used as a percentage of the average standard dosage (q=1.0 for a standard dose). It should be noted that the magnitude of the vector representing each herb's action depends upon the duration of the period over which we are considering its action. For the sake of consistency with future discussion, a standard period of two days is frequently used as the time over which a user consumes the standard dosage, a common convention used by many practitioners in America. Figure 2.1 is a graphical representation of an example of the preceding equation for three herbs.


Figure 1. Graphical representation of the combined action of three herbs. (The graph below portrays only two dimensions, though we really require 16 dimensions corresponding to the 16 TCM parameters in our model.)

Fig. 1


The preceding diagram illustrates the linear combination of effects from three herbs. If these herbs are chosen correctly, they will nudge the person toward a state of greater health (normalized physiological parameters, absence of TCM syndromes, etc.), represented by the origin point at the lower left of the graph. The case illustrated in Figure 1 is by far the most common situation in estimating the effects of combining herbs and drugs, and represents a linear summation of effects with no interactions. Side effects may nevertheless result if one uses herbal formulas that are contraindicated or overdoses on substances that all have similar actions.

Figure 2 illustrates another type of linear summation in which herb 2 is used to counteract some of the undesirable actions of herb 1, so that the net effect nudges the person toward the desired target point, the origin of the graph. Again, this is not an example of any interaction, but is a linear effect. An example would be the use of Rz Rhei for its purgative and Qi-moving actions in a case of Interior-Cold constipation, but adding a generous amount of Rx Aconiti Lateralis Preparata (zhi fu zi) and/or Rz Zingiberis (gan jiang) to create a net warming action.

Figure 2. Graphical representation of two herbs counteracting each other.

Fig. 2


True interactions

To account for the rarer non-linear interactions between pairs of herbs and/or drugs consumed at the same time, we modify Equation 2 to include interaction terms:

(Eqn. 3)     Eqn. 3

The vector functions fij represent the interaction between pairs of herbs i and j. If the relationship is strictly additive, fij is the null vector (all components equal zero). Theoretically, higher order interactions between groups of three, four, or more herbs may certainly be possible, but these types of interactions are rarely important in practice, and most herbal interactions can be understood by analyzing individual pairs.

Figures 3 and 4 illustrate two distinct types of true interactions, effects that could not be predicted based on the individual properties of each substance given by itself.

Figure 3. Graphical representation of two herbs interacting in a nonlinear manner, producing an enhancement effect.

Fig. 3


Figure 4. Graphical representation of two herbs interacting in a nonlinear manner, producing a qualitative change in the net effect.

Fig. 4


TCM herbalists have recognized that, occasionally, herbs interact with each other in strange ways. Sometimes they enhance each other's effects, such that the net effect is much greater than a simple addition of actions would predict. This possibility is illustrated in figure 3. Sometimes they interact to produce completely new actions, creating an action vector that points in a different direction than the simple summation of the two vectors. This possibility is portrayed in figure 4. Sometimes such an interaction may create toxic effects not present in either herb alone. This possibility is recognized by the medical profession as an undesirable "drug interaction". Alternatively, this type of interaction may create desirable effects in certain situations.


Synergies and antagonisms: current theories

It has been observed my herbalists and phytochemists that, in many cases, the presumed "active" ingredients of an herb, when extracted and isolated, do not have quite the same pharmacological effects as the whole plant preparation. That this is a common experience is often glibly explained as the result of complex chemical synergies within the whole herb preparation. Complexity in the number and type of chemical constituents may merely mean that the net effect of all together may be difficult to predict, and part of this difficulty may or may not be due to true synergies, or non-linear interactions between constituents. Perhaps the most convincing evidence of phytochemical synergies within a given plant species is discussed by James Duke. He points out various examples of pairs of chemical constituents having a measured pharmacological effect different from that expected from each constituent administered individually and discusses the theory that plants have evolved to make the most efficient use of their biochemical machinery. If clusters of closely related or complementary chemical constituents create adaptive synergistic effects, then such clusters will tend to be favored by evolution. If this theory is true, it weighs heavily against the interests of pharmaceutical companies in isolating a few active ingredients and then selling them to the public, if the whole raw herbs provide the advantages of inherent synergies of intra-species phytochemical groups.

Even assuming that synergistic effects among clusters of phytochemicals within a species might be common, if the net effect of such clusters is to maximize the adaptive pharmacologic effects for the survival of the plant itself, it does not necessarily follow that combining herbs together will result in even more synergies. There would not necessarily be any additional adaptive benefit for two plant species to evolve merely so that humans could gain synergistic effects when consuming them together. The only exceptions to this might be in plants that coexist in nature and have a symbiotic or commensal relationship. If such effects do exist, they are much more likely to be accidental.

Regarding potential interactions between drugs and herbs, one might argue that herbs and foods comprise a continuum of ingestible biological materials, and that if the risks of drug-herb interactions are so great, the very same thing applies to foods. In fact, this is often the case with especially dangerous drugs. Coumadin is known to interact with dietary patterns as well as herbs. MAO inhibitors are known to dangerously interact with many foods, including those containing tyramine, and some herbs: certain alcohol products, fava beans, certain cheeses, aged or processed fish, protein extracts, shrimp, yeasts, Rx Ginseng, Cornu Cervi Pentotrichum, Fructus Crataegi, and Radix Polygoni Multiflori. Drugs that have high likelihood of interactions with foods and herbs are inherently dangerous, because everyone has to eat food, and even common kitchen spices and herbs may have potent physiological actions. Such drugs, in principle, should receive greater scrutiny by the FDA, and less likelihood of approval. However, in the topsy-turvy world of FDA politics, it is often more likely that the blame will be shifted to the herbs, due to the tremendous political influence of the pharmaceutical companies.

In many cases, pharmacokinetic interactions, in which the ingestion of an herb or drug affects the rate of absorption or metabolism of other herbs or drugs, can be foreseen by understanding the properties of the herb in question, and in many cases should be obvious. Purgatives and cathartics, such as Rz Rhei and Fm Sennae, may interfere with absorption of drugs and nutrients from the intestine. Their purgative action will result in such drugs or nutrients being excreted; after all, the purpose of their action on the body is to flush the digestive tract. (This follows a corresponding TCM herbal principle that one should focus on either tonification or purgation, whichever is most appropriate at the moment; rarely are both done simultaneously.) Many drugs and herbal constituents are eventually degraded by detoxification systems in the liver; any actions to enhance these systems, such as in cases of Qi Stagnation of the Liver, would result in more rapid breakdown of such drugs or herbs and a reduction in effect of the standard dosage.

A commonly recommended method for avoiding or minimizing pharmacokinetic interactions is to arrange to ingest the herbs and the drugs at different times of day, or at least separated by many hours. This method may not completely avoid such interactions, especially in the case of herbs that affect liver detoxification rates.

Other common-sense considerations regarding possible side effects of combining herbs and drugs (not necessarily "interactions") would lead us to conclude that •[d1]• •[d2]•:

  • Stimulant herbs and Qi- or Yang-tonifying herbs, especially herbs that tend to increase blood pressure and level of mental and physical activity, should be used cautiously with drugs that also have these same actions: anti-depressive medications, stimulants, and sympathomimetics.
  • Herb with anti-inflammatory actions and which contain known steroidal compounds, such as Rx Glycyrrhizae (licorice), should be used cautiously with anti-inflammatory steroidal drugs, such as cortisone, prednisone, and prednisolone.
  • Herbs that enhance circulation and that may help invigorate Blood Stasis (myrrh, Rx Salvia Miltiorrhizae, garlic, ginkgo, etc.) should be either avoided entirely or used cautiously with drugs like Coumadin and heparin, which have similar actions but which are inherently highly dangerous, with or without herbs.

Most of the preceding issues involve a failure to consider the full effects of both the drug and herb in question, and are really issues of overdose on substance with similar actions rather than true interactions.



The net physiological and clinical effects of combining substances (herbs, drugs, foods, etc.) can be differentiated into linear, non-interactive effects, and interactive effects. True interactive effects are much less common than is reported in the media, since no attempt is made by popular media and even medical journals in defining what they mean when they use the term "herb-drug interaction". This term is often used to include incidents of overdose or inappropriate or contraindicated use, which are not, strictly speaking, unpredictable interaction, but merely linear summation of known effects of which the practitioner may have been ignorant.

It is a gross injustice to the herbal community to use the term "herb-drug interaction" to describe situations for which common ignorance, violation of contraindications, and other such misunderstandings are responsible. The term "herb-drug interaction" is often used by the media and the medical community to provoke fear and uncertainty regarding the possibility that unpredictable and dangerous interactions are responsible for many adverse events.

Whenever the term "herb-drug interaction" appears in the media and in government reports, the herbal community should demand that the term be explicitly defined. The results of such a demand may likely reveal that most of the people who use this term have little idea of what it really means.



Reference —{{ links }} will appear in a new window.