The enormous threat of AIDS to world health requires that all avenues be explored in the search for effective ways of controlling this disease. Recognizing that few therapeutic agents are effective in the treatment of AIDS and that these are often unavailable or too expensive to treat all those afflicted, WHO convened a consultation to determine how to investigate systematically traditional medicines that may have therapeutic benefit for many populations. While it may not be easy to apply current standards of drug Sevaluation to traditional remedies and natural products, investigation of these remedies may make some contribution to the relief of the current crisis. It is hoped that the guidelines for clinical trials presented below will be applicable under many circumstances and in many locations, and will encourage health care professionals to undertake drug trials that will uncover useful substances for improved treatments of AIDS and AIDS-related diseases. It is also hoped that they will encourage clinical studies that make accurate assessments and can be replicated in different populations.
2. PRECLINICAL CONSIDERATIONS
Clinical trials are expensive and carry some risk to those participating, either from adverse side effects or lack of information on the efficacy of the agent under study. Thus, it is of the utmost importance to identify clearly those approaches and agents that carry the highest potential for success before beginning a clinical trial. Some of the considerations that should be addressed before beginning a clinical trial are discussed below.
2.1 Reasonable assurance of potential human benefit either from prior clinical evidence or in vitro test results
There must be some prior evidence of the effectiveness of a potential agent against HIV infection and disease before it can be considered for a clinical trial. Such evidence might be anecdotal, concerning known clinical use, or come from pharmacological screening. Relevant pharmacological activity might include antiretroviral activity, modulation of the immune system, activity against opportunistic infections, or specific symptomatic relief.
Clinical effects that indicate the efficacy of a traditional medicine include the following:
(a) Decrease of symptoms such as diarrhoea, fever and pruritus, or improvement in appetite. Anecdotal reports of decreases in such symptoms should be interpreted with caution, however, as spontaneous remissions and exacerbations are apt to occur in the course of HIV infection. Greater certainty may be attached to such evidence, however, if symptoms go into remission for specific periods of time (e.g., fever or diarrhoea for at least one month) (see Annex 3).
(b) Disappearance of skin rash, Kaposi’s skin lesions, lymph-node regression or increase in weight.
(c) Clearing of opportunistic infections such as candidiasis, cryptosporidiosis, herpes simplex, cryptococcosis or toxoplasmosis.
(d) Improvement in stage of disease (proposed WHO staging system, Annex 3).
Laboratory indicators may also be used, although the value of these measures is less than that of clinical end-points. Some laboratory measure that have been used include p24 core antigen, beta-2 microglobulin, serum/urine neopterin, CD4+ lymphocytes, total lymphocytes or erythrocyte sedimentation rate (ESR).
A variety of in vitro assays are available for assessing anti-HIV activity. In vitro assays for antiviral activity should be performed in accordance with the procedures described in the report on in vitro screening for anti-HIV activity1 with special emphasis on cellular tests that monitor the complete viral replication cycle.
1Bulletin of the World Health Organization, 67(6): 613-618 (1989).
2.2 Botanical species identification
Plant species selected for trial must be identified by the currently accepted Latin binominal and synonyms, with associated authority, if known, along with the common names by which the plant is known. The part(s) of the plant used to make the preparation should be indicated, as well as detailed instructions for harvesting (e.g., stage of growth), storage and processing prior to and following formulation. The appropriate parts of the plant, in a state of sufficient integrity to allow for physical identification, should be provided to a qualified botanist.
2.3 Thorough literature search and analysis of findings
In order to be able to predict secondary pharmacological effects and potential toxicities of plants selected for clinical evaluation in HIV-infected subjects, or those with AIDS, the scientific literature must be thoroughly searched for the following types of information:
(a) biological effects of extracts of the plant(s) in the preparation being considered for human studies (including plants in closely related taxonomic groups);
(b) occurrence of secondary metabolites in the plants and their biological effects. Analogues of these substances should also be considered.
Adverse effects of some plant secondary metabolites may not be demonstrated in acute toxicological studies, but may be predicted to occur after long-term administration. For example, most pyrrolizidine alkaloids will cause liver toxicity, which can be life-threatening after prolonged administration. Risk-benefit ratios must be assessed and a decision made about using plants for clinical studies that have anti-HIV activity but also contain such known toxic substances as pyrrolizidine alkaloids.
The team involved in an anti-AIDS drug study can carry out literature searches and make its own reports if adequate library facilities are available, including access to on-line computer systems such as MEDLINE, TOXLINE, BIOSIS and NAPRALERT.
By special arrangement with WHO’s Traditional Medicine Programme, the WHO Collaborating Centre for Traditional Medicine, Chicago, will assist scientists, institutions and governments in developing countries by providing free access to the above-mentioned types of information from its NAPRALERT system.
Developing countries often have published information about local plants that is not abstracted or is unavailable in many developed countries; such information should not be neglected in the decision-making process. Established data banks are always pleased to receive such information.
2.4 Satisfaction of safety requirements
Pretrial toxicology requirements for all drugs are based on the proposed use of the drug and the information available regarding the interactions of the drug with biological systems. Traditional medicine products usually have a long history of use in humans. If acute side effects were severe and apparent, the use of the preparations would have been discontinued by traditional health practitioners. This hypothesis has been referred to as the “doctrine of reasonable assurance”. Thus, acute toxicological testing in animals may not be indicated. However, toxicological tests should still be performed to ensure that any previously unobserved long-term toxic effects are detected in animal models prior to the development of the protocol for a clinical trial.
When experience with a herbal medicine is too limited to make an assessment of long-term toxicity with confidence, additional toxicological testing should be undertaken before a clinical trial is initiated.
The primary purpose of the testing is to identify potential target organs and to include appropriate monitoring of the system subject to insult. However, in patients with HIV infection and disease, the presence of a variety of secondary disease states/opportunistic diseases may severely weaken organ systems, unmasking new organ susceptibilities. Animal testing may therefore be of limited value in predicting toxicity in HIV-infected persons.
The nature of the disease state being treated is also a prime consideration in determining toxicological testing requirements. HIV infection gradually produces a terminal disease for which there is thus far no effective treatment Every effort must therefore be made to test potential drugs promptly. A related consideration is that a traditional remedy intended for use in patients with advanced HIV disease may not require testing for long-term toxicity, since lengthy exposure to the drug will occur only in patients who are helped enough by it to sustain such an exposure. Chronic toxicity testing of such remedies may therefore be of dubious value. Extensive testing for long-term toxicity might better be reserved for herbal remedies that are considered for use in subjects in the relatively early stages of infection.
Toxicological tests in animal models should generally consist of a 90-day period of dosing by the route of administration intended for the clinical trial. Animals should then be necropsied and all major organ systems examined grossly and histopathologically. Such tests should be performed in two species, one of which is not a rodent. However, when adequate pharmacological or human experience exists, and national drug development policies do not require it, animal testing may not be necessary.
2.5 Adequate supply of the same plant material for all phases of the study
It is recognized that secondary plant constituents (including active principles) can vary both qualitatively and quantitatively in different lots of plant material. Therefore, it is essential that prelinicical testing and clinical studies be carried out on the same batch of plant material, collected at the same time from plants in the same general location. Enough material must be collected at one lime, bearing in mind that, after appropriate drying, fresh plant material can lose up to 90% of its weight.
An appropriate voucher specimen representing the collection must be prepared and identified by a qualified botanist and stored for future reference. Air-dried plant material must be milled to a coarse powder and stored in air-tight containers to prevent contamination.
After a dosage form has been selected (decoction, infusion, etc.), the extract to be used in the studies should be freshly prepared daily by a qualified pharmacist and should be properly labelled with instructions for use. In studies that may involve a placebo-treated group, some work may be required to create a dosage form that has similar organoleptic properties using a plant known to be devoid of in vitro anti-HIV activity.
3. PLANNING A CLINICAL TRIAL
Careful consideration must be given to the planning of clinical trials. The trial protocol must be written and approved in advance, in consultation with experts representing various disciplines. The following experts should be part of the team or available for consultation:
(a) Study coordinator - to ensure the feasibility of the study plan, i.e., that all patients can be followed up and that records on each patient can be completed according to the protocol; and to be responsible for the conduct of the clinical trial.
(b) Health care worker (physician) - to evaluate background evidence to determine whether a clinical trial is justified to assist in defining the study population and the clinical end-points that signify therapeutic success; and to help in the design of patient questionnaires.
(c) Nurse - to develop a plan for routine health care associated with the trial; and to assist the health care worker in designing the study and determining the feasibility of the protocol requirements, taking into consideration existing health care capabilities.
(d) Clinical pharmacologist - to evaluate background information including safety data to determine whether a clinical trial is justified; and to assist in preparing the study design.
(e) Pharmacist - to verify the formulation of plant substances; to prepare material for administration to the control group; and to distribute test and control drugs to appropriate patients in accordance with study protocol.
(f) Statistician - to estimate the size of study population needed in order to achieve statistically significant results; to plan for the analysis of study results at the same time the study is being designed; and to assist in the design of forms for collecting data and patient questionnaires.
(g) Traditional health practitioners - to assist in the preparation of the test drug in collaboration with the pharmacist.
(h) Social worker - to develop a plan for counselling patients during the study; and to assist in ensuring that social and behavioural considerations receive attention in the protocol.
(i) Others - additionally, access to a botanist and a natural products chemist on an ad hoc basis should be secured for the duration of the trial.
The trial protocol should define the objectives of the trial and give full details of the selection and allocation of subjects, treatment and treatment end-points, criteria for withdrawal from and termination of the trial, follow-up procedures, and data recording and analysis.
A simple, logical schema of the study objectives should serve as the introduction to the protocol.
3.2 Study sample
The sample population to be studied should be carefully defined.
The characteristics of subjects who qualify for the study should be summarized and include such information as age limits, sex, disease characteristics (e.g., severity, duration), and personal and social criteria that will ensure adequate participation in the trial (e.g., access to follow-up, ability to take medicine).
A similar list should be made of characteristics that disqualify a subject for the study, including other diseases and traits that may impair adequate participation in the study. Ordinarily, children and pregnant or lactating women are excluded from initial clinical trials.
The procedures to be used in recruiting subjects to the study should be outlined.
3.3 Allocation of subjects
Allocation of subjects to the various treatments included in the trial protocol should be made with due consideration to the following principles:
(a) Informed consent. Informed consent must be obtained from all study subjects according to all applicable standards of human rights and any relevant national regulations. Thus each subject must be given an adequate explanation of risk, benefit, and alternative treatments, and must be informed of his or her right to refuse to participate in the trial and, once in the trial, to withdraw from participation at any time.
(b) Randomization. Allocation should be randomized to avoid selection bias in the assignment of patients to treatment and control groups.
(c) Stratification, so as to achieve comparable control and treatment groups for purpose of analysis.
(d) Control group. An appropriate control group must be included as a basis for comparison with the treated group. Ordinarily, data on the control group are collected concurrently. Historical controls may be justified when the clinical course of any disease present in the treatment subjects is known to reach rapidly a clinically defined end-point. If there is no clearly beneficial treatment available, administering a placebo to the control group may be justified. Otherwise, the control group should receive the best alternative treatment available.
There are many social and behavioural aspects of HIV infection and AIDS which must be addressed when conducting a clinical trial. The trial treatment plan must be considered in the context of the environment in which the patient lives and is treated.
The dose of the drug and its schedule of administration must be chosen. These may be based on the customary dosage for other uses. However, dose titration may have to be conducted early in the investigation in a group of patients to establish a tolerated dose range. This is often referred to as a phase I study. Effectiveness is then evaluated using this dosage range in what is termed a phase II study.
To limit any potential toxicity from the drug, procedures must be developed to monitor signs and symptoms that may indicate toxicity. Certain forms of toxicity may be suspected from prior experience, but both clinical and laboratory evidence must be monitored throughout the trial to detect events indicative of unanticipated drug toxicity.
A course of action must be planned for dealing with toxicity if it occurs. In some cases, it may be sufficient simply to lower the dose and allow the patient to remain in the study. In other cases the patient’s participation in the study protocol may have to be discontinued.
3.5 Treatment end-points
Criteria for evaluating treatment must be standardized in advance. In the case of HIV infection and disease, however, the evaluation of any therapy is complicated by the large number of morbid events to which such patients are subject. A WHO staging system for HIV infection and disease has been proposed (see Annex 3)1 and because end-points other than death are difficult to define, it is suggested that measures such as a comparison of the percentage of patients who remain in their starting stage in both treatment and control groups during the course of the trial might be used. Additional information will be gained by enumerating all the morbid events listed in the staging classification that occur in the treatment and control groups.
1Weekly epidemiological record, 65, 221-228 (1990).
The sequence in the development of HIV-related clinical end-points and the time interval between them provides a clinical staging system which may be applicable only in a qualitative sense to patients in different communities. Thus such a general system may have to be modified to provide a specific instrument for the quantitative evaluation of individual clinical trials. The different spontaneous evolutions of the disease in different settings makes it essential that concurrent control groups are drawn from the same population in any therapeutic trial.
The stage of the disease should be assessed regularly and the trial continued long enough to determine whether the treatment has had any effect on the course of the disease. Because some of the evaluation parameters are subjective, bias may occur in the evaluation of patients. It may be crucial, therefore, to design the trial so that both patients and those gathering study outcomes are unaware of whether the patient has taken the study drug or is in the control group. Sometimes, however, this double-blind study design is not feasible. Under these circumstances a single-blind design may be used, in which either the patient or the health worker is unaware of whether the patient has been assigned to the treatment or the control group.
3.6 Criteria for withdrawing a patient from the trial
Criteria for withdrawing a patient from the trial, such as failure to adhere to the trial protocol, must be adopted before the trial begins.
3.7 Criteria for terminating the trial
Criteria for terminating the trial must also be adopted in advance. Two such criteria are: (a) unacceptable toxicity of the drug; (b) evidence of success or failure of therapy.
A schedule must be adopted to determine what information will be obtained at each follow-up visit. Such information should include a check-list of clinical symptoms (reported by the patient), clinical signs (reported by the health worker), and laboratory values. The optimal amount of information to be collected in a clinical trial may be difficult to determine. Gathering more information than absolutely necessary makes the trial cumbersome and may jeopardize its success. On the other hand, the systematic collection of data within the framework of a clinical trial may yield unanticipated data that prove valuable for further studies.
3.9 Data recording
An accurate record should be maintained for each patient during the trial. This record should describe the characteristics of each patient entering the trial, as well as the subsequent clinical course. Records of this kind facilitate the assembly of the basic data needed for any therapeutic trial. The record forms must be designed so as to provide space to enter all required data concisely to facilitate the complete, accurate and efficient collection of all necessary data. It is recommended that the following records be maintained for each patient:
(1) Enrolment record
(a) Has informed consent been obtained and documented?
(b) Basic patient information, e.g., date of birth, sex, and race.
(c) Check-list related to the present illness, which includes duration and a brief description of each relevant symptom.
(d) Check-list of required diagnostic tests.
(e) Treatment schedule.
(f) Check-list of other drugs taken.
(g) Check-list for exclusion criteria, e.g., childhood, pregnancy, lactation, and drug allergy.
(2) Medical history and physical examination record
(a) Review of symptoms, which might include a check-list as well as space for a brief description of any positive findings.
(b) Physical examination, which includes weight, a list of vital signs, and a check-list indicating which systems are normal and which are abnormal.
(3) Follow-up evaluation record
(a) Date of visit - visit number.
(b) Interim historical information.
(c) Medication compliance evaluation.
(d) A recording of adverse events, including a check-list of those anticipated, as well as other events that may indicate serious unanticipated toxicity.
(e) A list of any concurrent drug therapy, including self-medication.
(f) Clinical evaluation, which includes height, weight, vital signs, and specific signs and symptoms of the disease.
(4) Laboratory results
Laboratory results should be recorded on a form that lists the date together with values for such data as haematology, blood chemistry, and urine analysis that are required for the trial.
Individual records must be checked to ensure that they are complete. Then data from individual records must be compiled for analysis.
3.10 Data analysis
Statistical aspects of the trial must be considered in advance. An estimate must be made of the number of patients needed in the study to make it likely that a statistically significant result will be attained if the treatment is as effective as anticipated. Ordinarily, a plan for data analysis should be determined in advance, along with any necessary rules for stopping the trial, if the anticipated results of the trial are obtained earlier than expected.
Analysis of trial results may be carried out on an interim basis to assure that no unanticipated toxic events are occurring and to ensure that the trial is not continuing beyond the point needed to achieve the study objectives. Sometimes it may be necessary to have an independent committee to monitor interim results. However, when all results have been obtained, the analysis should be carried out in the manner that was specified when the trial was designed. In addition, however, a great deal may be learned by examining the data to determine whether they suggest unanticipated results that may have therapeutic or scientific interest or that may suggest hypotheses for future studies.
Analysis must also be made of data on all patients who have left the trial, including their fate and the reasons they left the trial. An effort must be made to trace every such patient.