My Medical Philosophy: Formula-Composition Rules for the TCM Large-Formula Multi-Target Special Therapy

How does one compose a large formula that may work well? Are there referenceable principles for doing so?

The basic principle of traditional TCM formula composition is to combine herbs according to the "seven relationships" (qi qing) first recorded in Shennong Bencao Jing (Divine Farmer's Materia Medica). The "seven relationships" are the seven ways Chinese herbs relate to one another: single use (danning), mutual reinforcement (xiangxu), mutual assistance (xiangshi), mutual restraint (xiangwei), mutual detoxification (xianggsha), mutual antagonism (xiang'e), and mutual incompatibility (xiangfan).

"Single use" means one herb works alone without assistance; TCM has many such single-herb formulas. For example, Dushen Tang (Lone Ginseng Decoction, made from just Renshen/Ginseng) strongly greatly supplements original qi; Dusheng San (Lone Sage Powder, made from just Puhuang/Typha) stops bleeding; and Gancao Tang (Licorice Decoction, made from just Gancao/Glycyrrhiza) treats sore throat.

"Mutual reinforcement" means combining two herbs with similar effects so they enhance each other. For example, Mahuang (Ephedra) plus Guizhi (Cinnamomum) strengthens sweat-release and exterior-resolving; Shigao (Gypsum) plus Zhimu (Anemarrhena) enhances heat-clearing and fire-draining.

"Mutual assistance" means one herb is the chief and another the assistant, with the assistant boosting the chief's effect. For example, Huangqi (Astragalus) is often paired with Fuling (Poria), which increases Huangqi's qi-tonifying and diuretic action.

"Mutual restraint" means one herb's toxicity or side effects are checked by another. For example, Shengjiang (Fresh Ginger) reduces the toxicity of Banxia (Pinellia) and Fuzi (Aconite); honey reduces the toxicity of Wutou (Aconiti Tuber). This is called "Banxia fears Shengjiang" and "Wutou fears honey."

"Mutual detoxification" means one herb can eliminate another's toxicity. Xiangwei and xianggsha are bidirectional: "Banxia fears Shengjiang" equals "Shengjiang detoxifies Banxia."

"Mutual antagonism" means that when two herbs are combined, one weakens the other's original effect. For example, Laifuzi (Raphanus seed) has a strong diuretic action and accelerates the metabolism of Renshen (Ginseng), weakening its qi-tonifying effect. When using Renshen to tonify qi, one should avoid Laifuzi. On the other hand, mutually antagonistic herbs can also counteract side effects. For example, after taking Renshen, if "ginseng abuse syndrome" appears (high fever, constipation, flushed face and ears, restlessness), using Laifuzi or raw radish quickly relieves the symptoms.

"Mutual incompatibility" means that two herbs used together produce severe toxicity and serious side effects; these are compatibility contraindications. TCM has the "eighteen incompatibilities" and "nineteen mutual fears," all compatibility contraindications. But viewed today, these contraindications are unscientific; some were simply fabricated by ancients out of conjecture. For instance, the ancients believed white and black were water-and-fire incompatible, so white and black herbs were deemed antagonistic—which is absurd. At the same time, successive generations of TCM also discovered other toxic combinations not listed in the eighteen; for example, combining Yuanhu (Corydalis) with Maqianzi (Strychnos) increases toxicity. I once summarized herbal-toxicity knowledge in a long article on this topic; interested readers can find it in my older works; I will not repeat it here.

Overall: to enhance efficacy, use combinations with mutual reinforcement and mutual assistance; to reduce a herb's toxicity, consider combinations with mutual restraint and mutual detoxification. Antagonistic herbs, though not causing severe side effects, reduce efficacy, so use them sparingly—only to mitigate side effects. Herbs of opposite properties must be prohibited.

Looking at these rules today, they are quite reasonable principles of drug combination. Even now, whether with Chinese or Western drugs, our basic combination principles largely follow these rules.

In building a large formula, we should fundamentally follow these same principles. But the more kinds of drugs we command, the harder those principles are to apply. The actions, side effects, and interactions among drugs are all complex. This is why large formulas have long been criticized by many doctors—and with reason.

But, as I argued in earlier articles, complex disease problems cannot be solved by simple thinking. When we face a patient whose cause and pathology are deeply complex, we must consider all the etiological factors; too few herbs cannot address all the patient's problems.

The patient's various problems also interact. If we solve one but not the others, the moment we stop the drug, the unsolved problems trigger a relapse of the solved one, and all our previous effort is wasted.

A large formula is therefore a possible solution for refractory disease. Only by tackling interrelated problems in parallel can we comprehensively resolve the patient's condition and reduce rebound and relapse.

For easy-to-treat diseases, all this is unnecessary. For common minor ailments I myself prefer simple formulas or small combinations of just a few herbs; I do not like large prescriptions. Such small formulas are much simpler to compose but mostly address self-limiting conditions.

A small formula may occasionally cure one or two refractory cases, but repeatable results are rare, and one cannot expect a high effective rate. Treating refractory disease this way easily wastes the patient's time and misses the precious therapeutic window.

When I first began composing large formulas, I simply added several formulas of some efficacy together. The reasoning was simple: if one possibly effective formula has a 1% success rate, two used together might have 2%. Facts I observed in clinical practice confirmed that such synergy sometimes occurs.

For example, my strong antiemetic formula (published in an earlier article and to be revisited in the specific-treatment section) combines several common vomiting formulas—Xuanfu Daizhe Tang, Erchen Tang, Jupi Zhuru Tang, Dingxiang Shidi San. In treating refractory vomiting in chemotherapy patients, it outperforms almost all other antiemetics.

This seemingly clumsy method is actually the most efficient. Because these formulas have been used repeatedly through history, their efficacy is verified countless times. If combining herbs within a formula can enhance effect, then by the same logic combining formulas can enhance effect too.

Many people worry that in a large formula, herbs are prone to chemical reactions. That is possible, but from a modern chemistry standpoint, a single Chinese herb already contains hundreds of chemical constituents, sometimes more. Mixing several herbs in a small formula can equally produce complex reactions; a large formula only raises the probability slightly.

And, as we know from biochemistry, biochemical reactions do not happen easily. Many require enzyme catalysis, and enzymatic reactions are highly specific under stringent conditions. So multi-drug combination is not as risky as imagined.

We can treat a simple formula as a single herb, just with more complex constituents. Combining formulas follows the same principles as combining herbs—the seven relationships: single use, mutual reinforcement, mutual assistance, mutual restraint, mutual detoxification, mutual antagonism, and mutual incompatibility.

Beyond the seven relationships, TCM composition has other principles, such as jun-chen-zuo-shi (sovereign-minister-assistant-courier). But as for the jun-chen-zuo-shi theory, I personally believe grasping its spirit is enough; do not be bound by the rigid rules the ancients devised.

The principle of jun-chen-zuo-shi is not essentially different from the seven relationships; it is just another formulation. But later literati, over-egging the pudding, added many rigid and clearly unreasonable rules. For instance, Suwen says: "One sovereign with two ministers is a small formulation; one sovereign with three ministers and five assistants is a medium formulation; one sovereign with three ministers and nine assistants is a large formulation." This becomes a word-and-number game that severely shackles the clinician's prescribing mind.

I want to use modern mathematical thinking to further explain the large-formula multi-target approach I advocate. But I ask readers not to treat these as fixed rules; master the way of thinking and design your own approach based on the facts.

We can borrow math concepts like "like terms" and "factors" to understand large-formula multi-target prescribing. Similar formulas are like like terms: they can be combined to focus on one aspect of the patient's problem. For instance, for ascites in advanced cancer, we can merge historically effective empirical formulas for ascites into one like-term group.

Formulas with different actions but all useful for the same patient can be treated as different factors, combined by multiplication. For example, advanced cancer may involve not only ascites but also cancer pain and cancer fever. Formulas for ascites form one like-term group; those for cancer pain another; those for cancer fever a third. These groups can then be multiplied together.

With these two mathematical concepts, any reader with junior-high math can grasp it at a glance. I explain it this way not to force math and TCM together, but to help readers understand the composition approach more simply.

Math and medicine are different domains; they have parallels and differences. Medicine values practice, and this section uses induction and deduction. As everyone knows, conclusions from induction and deduction may be right or wrong; their truth must be verified by practice (animal and human experiments).

Practice requires statistics, so I especially hope this method can be combined with animal experiments to verify efficacy and safety statistically. Exploring higher-effectiveness regimens in animal studies before applying them clinically is the safer route. It is a pity that most TCM physicians lack such conditions.

We can also study this prescribing approach with controlled trials. In treating patients, I used controlled studies to continually refine formulas built on the large-formula multi-target idea.

I first set Formula A and Formula B, where B is A plus one herb or one group. Similar patients were split: some got A, others B. If B outperformed A, all subsequent patients got B first. Then I added another herb or group to B to make C, tested on another group; if C outperformed B, C became the preferred regimen. And so on. I repeated this for about ten years.

I hope colleagues who learn large-formula multi-target prescribing will similarly keep refining their formulas in treating various diseases, improving safety and efficacy. If we ran such controlled experiments in the lab, we might find better regimens faster. Such research may require national funding and large research institutions; whether that opportunity comes, we leave to fate.