Essentials of Complementary and Alternative Medicine (June 1999)

Components of Safety
 
Mislabeling
 
Misrepresentation
 
Comparative Trials
 
Misapplication
 
Misdiagnosis
Definitional Issues
Indirect Risks
Adverse Effects: Types A, B, C, and D Reactions
Limitations of Traditional Experience
 
Nontraditional Hazards
Product-Related Determinants of Adverse Effects
Tools for Safety Assessment in Complementary Medicine
 
Clinical Accuracy
 
Methods of Measurement
 
Rare Events and Public Health Impact
 
Types of Evidence in Adverse Effects Assessment
The Safety of Selected Complementary Practices
 
Poison Control Centers
 
Randomized Controlled Trials
Summary
Chapter References
C
OMPLEMENTARY AND ALTERNATIVE MEDICAL (CAM) PRACTICES ARE INCREASING IN POPULARITY. IN THE UNITED STATES IT IS ESTIMATED THAT AT LEAST ONE-THIRD OF THE 
POPULATION USES THESE PRACTICES ON A REGULAR BASIS, CONSTITUTING 425 MILLION VISITS AND SPENDING ALMOST $14 BILLION ON THESE PRODUCTS AND PRACTICES 
ANNUALLY; MOST OF THESE COSTS ($10 BILLION) ARE OUT-OF-POCKET EXPENSES (
1
). IN EUROPE, THE PROPORTION IS EVEN HIGHER, WITH 40 TO 70% OF THE POPULATION 
USING CAM IN SOME CASES (
2
). THE WORLD HEALTH ORGANIZATION (WHO) ESTIMATES THAT 80% OF THE DEVELOPING WORLD'S POPULATION DEPENDS ON TRADITIONAL 
PRACTICES (MOSTLY HERBAL PREPARATIONS) FOR THEIR HEALTH CARE MANAGEMENT (
3
). IN 1996, A STUDY IN THE LANCET EXTRAPOLATED ON THE BASIS OF ABOUT 3000 
PERSONAL INTERVIEWS WITH SOUTH AUSTRALIAN PERSONS AGED 15 YEARS OR OLDER THAT THE AUSTRALIAN POPULATION HAD SPENT $AU621 MILLION FOR ALTERNATIVE 
MEDICINES AND $AU309 MILLION FOR ALTERNATIVE THERAPISTS, COMPARED TO $AU360 MILLION OF PATIENT CONTRIBUTIONS FOR THE CONVENTIONAL DRUGS THAT HAD 
BEEN PURCHASED IN 1992–1993 (
4
).
As use of these practices rises, concern over safety becomes increasingly important. Toxic products can be used in a relatively safe manner if they are delivered by a 
well-trained practitioner with supporting medical infrastructure and close monitoring. However, many practices used in CAM involve self-care or delivery by 
practitioners whose training may be unknown, unmonitored, or inadequate (
5
, 
6
 and 
7
). Because there is an increasing potential for public harm with misuse of these 
practices, closer scrutiny of potential adverse effects from CAM practices is needed.
On the supply side, the growth of the market for complementary medicines is promoted by active marketing, which regularly emphasizes the natural origin and 
harmlessness of these medicines (
8
, 
9
, 
10
 and 
11
). On the demand side (i.e., the market), the volume is increased by various factors, which vary from user to user 
and from product to product (
12
, 
13
). Some consumers use alternative medicines because of the health-promoting action ascribed to these products (
8
, 
14
), whereas 
others take them because conventional medicine does not offer an adequate prospect of being cured (
15
, 
16
 and 
17
). The preference for alternative medicines may 
be partly explained by the consumer wanting to exert control over his or her own health status; also, the consumer may be attracted to the patient-oriented approach 
of the alternative practitioner (
17
, 
18
, 
19
 and 
20
). In addition, advice of family or friends can be a primary or additional motive to start the use of alternative products 
(
21
, 
22
). Finally, many conventional health care providers are increasingly tolerant of rather than rejecting toward alternative therapies (
23
, 
24
 and 
25
).
BROAD SCOPE
Because of the broad scope of CAM, it is difficult to define what preparations should be discussed and what should be excluded in this review. The same product may 
be an officially licensed medicine in one country and a dietary supplement in the next, and it is also possible that the same substance is available as an approved 
medicine and as a health food preparation in the same country. Furthermore, current markets of dietary supplements and health food products not only comprise 
medicinal preparations, but also products used for other purposes, such as aphrodisiacs (
26
), mind-altering recreational herbs, and preparations for the enhancement 
of athletic performance (
27
, 
28
 and 
29
). Although these types of products may be outside the domain of complementary medicines in a strict sense, we have included 
them here because they are often available through similar outlets as complementary medicines and are often perceived by the public as similar products. Also, we 
felt that the informational value of this chapter would benefit from a broad scope rather than a conceptually strict definition.
THE RELATIVE NATURE OF “SAFETY”
Safety is defined by Webster's Dictionary as “freedom from whatever exposes one to danger or from liability to cause danger or harm; safeness; hence the quality of 
making safe or secure or of giving confidence, justifying trust, insuring against harm or loss, etc.” It derives from the Sanskrit word  sarva, which is akin to the word 
salus, meaning “of sound condition, well preserved, unharmed, whole.” The word  safety is thus intimately related to the concepts of health and wholeness, or 
completeness, and refers more specifically to conditions that threaten wholeness or health.
The medical application of the term safety is relative. This is illustrated by the fact that the United States Food and Drug Administration (FDA), which is charged with 
ensuring safety of products and devices, has no specific definition of the term. Safety is relative to several factors, including:
1. Toxicity of the product or device being used.
2. Potential benefits incurred from use.
3. Context of use—e.g., self-care or under the care of a competently trained professional with knowledge, skills, and abilities to ensure proper use.
4. Appropriate monitoring, marketing, and advertising to ensure that its use is accessed properly.
5. Values that underlie judgment of its proper use—for example, the desired and expected outcomes interpreted by judgments that determine whether its effects 
are adverse or beneficial.
To speak of safety, one must address the risk-to-benefit ratio derived from engaging in these practices or administering products with therapeutic intent. The 
risk-benefit ratio becomes the index by which safety is judged. For most products or practices, only the direct assessment of the risk-benefit ratio can provide reliable 
information with which to make evidence-based judgments about safety.
THE RELATIVE SAFETY OF “NATURAL”
CAM products are often assumed to be safer than conventional medicine and drug therapy, providing a frequent reason individuals turn to CAM (
30
). The term natural 
medicine, used frequently in Germany and elsewhere, attracts people to CAM practices who may be afraid of the side effects of conventional therapies. People may 
assume that therapies labeled natural are inherently safer, which is not always true.
Some CAM practices appear to be generally of low risk. Examples include meditation, mind-body techniques (e.g., visualization), psychic or spiritual healing, and 
whole systems of practice, such as homeopathy and acupuncture. Other complementary therapies clearly present direct risks. Examples are herbal remedies; 
intravenous hydrogen peroxide or high-dose megavitamin and mineral infusions; and procedures such as colonics or high-velocity spinal manipulation. Yet, given the 

risk-benefit definition of safety, one cannot assume any therapy, which on the surface appears mostly harmless, is safe.
Homeopathy provides an example of the need to match appropriate criteria to specific CAM therapies. Homeopathy, which uses highly dilute substances, appears to 
be inherently safe. Homeopathic remedies often begin with highly toxic substances. But, through a process of serial dilution, these substances end up having little or 
none of the toxic component remaining in the final preparation. However, specific precautions are still required. For example, not all homeopathy preparations (or 
those labeled as homeopathic) are extremely dilute. Some “low-potency” preparations or those mixed with herbal products may contain potentially dangerous 
substances (
31
).
One may also assume that prayer has no direct toxic effect. However, adverse effects from “psychic” phenomena have been documented (
32
). Whether one 
understands the mechanisms (e.g., psychogenic, cultural, autonomic), can we reasonably assume that these practices have no adverse effects?
Therapies with known toxic effects (e.g., herbal preparations) may actually prove to be inherently safer (given their relative benefits) than comparable conventional 
therapies. An example is the use of hypericum, or St. John's Wort, for the treatment of mild to moderate depression. The efficacy of this herb for this indication has 
been demonstrated in clinical trials and has very few adverse effects. In direct comparison with conventional antidepressant agents, hypericum has fewer side effects 
with comparable efficacy (
33
). This finding illustrates that a therapy with higher potential direct toxic effects than homeopathy or prayer may be inherently more 
desirable—that is, “safer”—given the indications and comparative value to a conventional therapy because of a reduced risk-benefit ratio. Therefore, decisions about 
safety requires evaluation using various criteria, both for systems as a whole (e.g., homeopathy, prayer, herbalism) and for specific products within systems during 
application for specific indications.
To be considered sufficiently safe for application, certain therapeutic systems (e.g., acupuncture and homeopathy) may require only toxicity screening and appropriate 
training for their use as a whole system. In this case, specific product safety testing for each indication is unrealistic. Other inventions (e.g., herbalism or megavitamin 
therapy) require screening of both the entire system for general safety characteristics and specific testing of standardized products for specific indications before 
marketing. Examples of these approaches are the WHO  Guidelines for Safe Acupuncture Treatment (
3
) and ongoing work in several countries developing both safe 
and practical regulation of botanical products (
34
, 
35
 and 
36
). (See 
Chapter 5
 for guidelines on how to use potential risk in making evidence-based decisions about 
CAM.)
COMPONENTS OF SAFETY
Adverse effects from CAM practices can be classified into three broad categories:
Direct adverse effects are either short-term, in which they are often called  toxic effects, or long-term, in which they are called side effects.
Indirect adverse effects are those events that occur because of incompetent delivery of the therapy or diagnostic procedure.
Definitional adverse effects. Many CAM systems use different diagnostic categories, patient preferences, explanatory models, and outcome values than those 
commonly accepted in the West. Failure to provide clarity in these definitional and descriptional areas can lead to misunderstanding of application or attempted 
application of the practice, which can also produce adverse effects.
Mislabeling
Mislabeling occurs when a product or device does not contain the items it purports to contain or does not perform the actions its claims to perform. In products such 
as herbal preparations, mislabeling may involve the application of a specific herbal name and yet the product may not contain that herb. For example, confusion and 
mislabeling of various ginseng species has lead to inaccurate claims of toxicity. Herbal products sometimes are labeled as containing a single herb, but actually 
contain products from a variety of plants (
37
). Some homeopathic products are labeled homeopathic because they include highly dilute preparations combined with 
plant products, or even pharmacological doses of conventional drugs. Likewise, acupuncture needles may not be prepared with appropriate manufacturing and 
sterility standards and yet be labeled sterile. Mislabeling may occur because good manufacturing processes are not incorporated in the manufacture of the product. In 
fraudulent cases, mislabeling may occur deliberately. Although fraudulent mislabeling of products occurs, its frequency is unknown and probably represents a minority 
of products compared to those who mislabel because of inadequate manufacturing and standardization procedures.
A significant subset of mislabeling is underlabeling. Underlabeling occurs when potentially toxic or active ingredients are not listed on the product's contents label. For 
example, some Traditional Chinese and Ayurvedic herbal preparations may contain toxic doses of mercury, lead, arsenic, and other heavy metals (
38
). In some cases, 
these substances are in the product because practitioners in the CAM system believes they provide important therapeutic effects. These substances are clearly toxic 
and have long-term damaging effects, especially if ingested by children, but may not appear on the label. Nor can the predictions about their concentration be 
estimated from the name of the plant product or herbal mixture. The recent Dietary Supplement Health and Education Act passed by the U.S. government may have 
aggravated this problem by allowing some products that are frequently used as therapeutic agents to be classified as foods. This classification results in less labeling 
of the content and description of appropriate usage. Underlabeling can be a serious type of mislabeling.
Misrepresentation
Misrepresentation involves claiming effectiveness for ineffective interventions or diagnostic procedures. Even in relatively benign conditions, misrepresentation can 
cause adverse effects because of the risk of wasting time and money, the increased use of diagnostic techniques (including repeated history-taking and physical 
exams), and the generation of false expectations about the outcome. In addition, misrepresentation of ineffective therapies causes harm if effective therapies are not 
used for the alleviation of the condition.
There are two key decisions needed to evaluate misrepresentation. The first is whether an intervention is needed at all. Many conditions have a benign natural course 
or recover spontaneously and require no intervention. In these cases, any intervention carries potential risks. For example, close to 80% of individuals who visit a 
general practice in which a specific diagnosis is not reached will have his or her problem resolve or improve with no therapy (
39
). In this case, any intervention with 
potential side effects, inconvenience, cost, or one that produces anxiety or other psychological factors is inappropriate. Thus, prognostic need is an important part of 
evaluating misrepresentation. Increased self-care with natural medicines and the assumption of comprehensive medical care by practitioners without extensive 
experience and training in the management of serious disease can lead to unfavorable, avoidable, and, therefore, adverse outcomes. Examples of these outcomes 
have been published, such as an attempted treatment of diabetes mellitus using ineffective herbal preparations (
5
, 
6
 and 
7
). Other examples include the treatment of 
cancers in the early stages using biological or homeopathic preparations and the reliance on prayer and mental healing while serious conditions progress (
40
). The 
prevalence of such events from a public health perspective is unknown.
The first aspect of misrepresentation can only occur if there is truly an effective conventional therapy available for the condition. In many incidences, there may not be 
an effective therapy. For example, acupuncture or homeopathic treatment may alleviate pain and improve function in patients with pancreatic cancer because of 
nonspecific effects. Because the side effects of these therapies are less than those produced by chemotherapeutics used in conventional medicine for the treatment 
of pancreatic cancer, and because conventional treatment does not change the prognosis of this disease, the use of these two alternative modalities in the treatment 
of pancreatic cancer could not be considered as causing harm or misrepresentation in this sense.
Comparative Trials
Direct comparative trials of CAM therapies compared with proven conventional therapies has increased. There are now a number of examples comparing 
acupuncture, homeopathy, and herbal preparations, as well as spinal manipulation and Traditional Chinese Medicine with and without the incorporation of 
conventional medicine in randomized controlled trials. These trials often, but not always, demonstrate similar benefit with equal or reduced adverse effects from 
complementary therapies, although costs are rarely measured.
Examples of such direct comparison is the hypericum versus imipramine trial described previously (
41
), the comparison of classical homeopathy versus salicylates in 
the treatment of rheumatoid arthritis published by Gibson et al. (
42
), and the controlled trial comparing antacids, cimetidine, deglycyrrhizinated licorice, and gefarnate 
in the treatment of chronic duodenal ulceration (
43
). Hammerschlag has summarized the direct comparative trials of acupuncture versus conventional medicine in a 
variety of conditions. In most trials in which the benefit of conventional treatments is similar to acupuncture, adverse effects were usually equal or less (
44
). More trials 
of this type are needed in complementary medicine.

Misapplication
Misapplication of effective therapies can also result in harm. Misapplication usually occurs because of inadequate training, in knowledge, skills, or experience or from 
incompetence or practitioner impairment. Conventional medication has well-established training, certification, licensing, and monitoring procedures to ensure that the 
knowledge, skills, and qualifications of a practitioner are adequate. Even with these safeguards, misapplication occurs. If CAM practitioners engage in the entire 
scope of a medical practice, then their knowledge, skills, and experience in these areas must be at least comparable to those involved in conventional primary care, or 
their practice should be specifically restricted and defined. Appropriate referral to and from medically qualified practitioners requires knowledge by those practitioners 
of which conditions are best managed and under what circumstances. As the number, type, and background of complementary practitioners increases and as the 
identification of more effective and ineffective CAM therapies occurs, it is likely that a generalist medical coordinator with knowledge about both CAM and 
conventional medicine will be needed to prevent misapplication.
Misapplication can occur if practitioners fail to properly refer to or apply effective practices. Schools of acupuncture for physicians consist of about 200 to 400 hours of 
training in acupuncture, which is often considered inadequate in countries where extensive use and training of traditional acupuncturists has occurred for centuries. In 
addition, non-MD acupuncturists may have inadequate medical training. The WHO is attempting to address these problems by creating categories of acupuncturists 
with various training standards and scopes of practice. Thus, minimum hours and standards for the training of qualified medical doctors in acupuncture is defined, as 
are minimal training and experience in anatomy, physiology, pathophysiology, prognosis, and scope of practice for acupuncturists in Western diagnosis and medicine. 
Without such standards for all those who may engage in, refer to, or recommend complementary practices, misapplication is likely to continue unchecked. In the 
United States, there are at least three certifying bodies in acupuncture and Oriental medicine, each having different requirements for certification and scopes of 
practice. States seeking to provide appropriate laws for licensing and regulating acupuncturists may use criteria from different organizations or from none of these 
organizations in their certification process. A similar situation exists with homeopathic practitioners, having various certification and licensing and training 
requirements for MDs, NDs, DCs, and even lay practitioners. Whereas all 50 states recognize, license, and regulate chiropractors, only about 24 states have statutory 
regulations dealing with acupuncture, 12 with naturopathy, and 4 with homeopathy, one of which dates back to 1906. The situation in the United Kingdom is even 
more chaotic, where the practice of medicine does not require any certification or licensing. (See 
Chapter 2
 for a list of current licensing laws for various practices by 
state.)
Misdiagnosis
Finally, harm can occur because of incorrect, inaccurate, or inappropriate diagnosis and patient classification, resulting in the application of effective therapies for the 
incorrect condition. Harm can also occur from failure to detect conditions that can be corrected. Harm from inappropriate diagnosis and patient classification is 
different from suboptimal application of effective therapies. The former involves clinical disagreement and diagnostic accuracy, or failure to detect and apply the 
therapy to the main cause of a condition. The practitioner should especially be alert for CAM systems that create their own diagnostic class for which their therapy is 
then applied. Iridology claims to diagnose specific conditions from patterns in the eye. This results in unnecessary treatments not proven to effect that particular 
condition (
45
). Many other similar diagnostic methods are available that in principle set up situations for the same risk. Electrodermal diagnostic methods, for example, 
claim to be able to detect functional changes in acupuncture meridians and subsequently the corresponding organ locations. These measurements then become an 
“electromagnetic” diagnosis of disease. These “diseases” created with such instruments are then “treated” using a variety of means. Without proof that such a 
“diagnosis” is important, these treatments are claimed to be “preventive,” and patients may go through multiple “measurements” and “treatments.” This can cause 
harm by creating anxiety, excessive and unnecessary testing, treatments, and expenditures of time and energy, all of which have unknown benefit. Similar CAM 
diagnostic systems, including combinations of endocrinological and blood measurements from normal serum, radiathesia, intuitive diagnosis, auricular diagnosis, and 
the use of pleomorphism, all carry the same risk of misinformed application and misdiagnosis.
Orthodox medicine has its share of similar questionable screening and diagnostic methods, including tests such as PSA for all men, mammography for younger 
women, occult stool blood tests, fetal monitoring in normal pregnancies, and so on. In many of these tests, their reliability and correlation with significant outcomes are 
questionable, yet their application leads to interventions, all of which carry risks. An overdependence on medically “correct” (e.g., objective, proven, most common, 
most easily verified, most high-tech) outcomes in place of attending to subjective, quality of life, and individually valued outcomes is probably a major reason why the 
people turn to CAM practitioners (
30
). These are definitional and value issues that if not addressed can lead to adverse effects from treatment.