NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Hardy M, Coulter I, Venuturupalli S, et al. Ayurvedic Interventions for Diabetes Mellitus: A Systematic Review. Rockville (MD): Agency for Healthcare Research and Quality (US); 2001 Sep. (Evidence Reports/Technology Assessments, No. 41.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Ayurvedic Interventions for Diabetes Mellitus

Ayurvedic Interventions for Diabetes Mellitus: A Systematic Review.

Show details


We synthesized evidence from the scientific literature on the effectiveness of Ayurvedic therapies for diabetes mellitus using the evidence review and synthesis methods of the Southern California Evidence-based Practice Center (SCEPC). This is one of the designated centers established by the Agency for Healthcare Research and Quality for the systematic review of literature on the evidence for benefits and harms of health care interventions. The project staff collaborated with the National Center for Complementary and Alternative Medicine, with the project officer at AHRQ, and with a group of technical experts representing diverse disciplines.

Scope of Work

Our literature review process consisted of the following steps:

  • Identify sources in the literature reporting evidence for Ayurvedic therapy.
  • Develop a strategy to maximize retrieval of Ayurvedic literature.
  • Conduct a search of the Ayurvedic (broadly defined) literature to identify topic areas with sufficient publications to support a detailed review.
  • Develop a strategy to identify English-language literature published in India but not normally cited in the Western indexes.
  • Conduct a focused literature search of therapies for the treatment of diabetes in the Ayurvedic literature from both Indian and Western sources.
  • Assess the strategies for completeness.
  • Evaluate articles for methodological quality and relevance.
  • Extract characteristics and results from studies meeting methodologic and clinical criteria
  • Synthesize the results.
  • Submit the results to technical experts for review.
  • Incorporate the reviewers' comments into a final report for submission to AHRQ.


We conducted a literature search of the field of Ayurveda to establish the distribution of studies using Ayurvedic interventions. These studies were then evaluated to determine if there was a sufficient body of literature in any one combination of disease and/or Ayurvedic modality to enable a comprehensive systematic review.

The initial search was guided by the following research questions:

  • Is Ayurveda studied as a whole system?
  • What disease states or modalities are the major focus of the Ayurvedic studies?
  • Is there a sufficient body of literature in any one combination of disease and/or modality to do a traditional systematic review?
  • If not, what type of review is possible?
  • Is it possible to access the literature in India in a manner that is both cost-effective and comprehensive?
  • Is the Indian body of literature qualitatively different from Western literature?

Based on the results of this search and on further discussions with our technical experts and the sponsoring agencies, we chose Ayurvedic therapies for diabetes mellitus as the focus of the comprehensive review presented in this report.

Ayurvedic Literature Search Design

Technical Expert Panel

We recruited a group of technical experts to advise us. The technical experts were from diverse disciplines including acupuncture, Ayurvedic medicine, chiropractic, dentistry, general internal medicine, gastroenterology, endocrinology, integrative medicine, neurophysiology, nursing, pharmacology, psychiatry, psychoneuroimmunology, psychology, sociology, and traditional Chinese medicine. These experts assisted the project in several ways. They helped the research group in defining and conducting the initial overall survey of the field of Ayurvedic medical research; they reviewed the results for completeness; and they were consulted on what topics appeared to be good candidates for a comprehensive review. Members of the expert panel reviewed the search terms we used; some members also acted as reviewers of the preliminary report along with the reviewers listed in the acknowledgements. Reviewers and members of the expert panel and their affiliations are listed in Appendix B.

Preliminary Search of the Ayurvedic Literature

We conducted a preliminary and nonsystematic search of Western databases to quickly survey the Ayurvedic literature and to help us define our search strategies. To determine what might be available in the Ayurvedic literature, we performed a preliminary search in PubMed, an online database encompassing the Cochrane and MEDLINE® databases from 1966 to 1999. This simple review yielded only 708 citations. Because of the low yield, we then tested the strategy of surveying all the Asian Indian journals indexed in the Index Medicus from 1966 to see how many additional published studies these sources yielded. There were fewer than 1,200 citations published in these journals, and the majority of them covered conventional Western medical topics. Most of the studies that appeared to focus on Ayurvedic medicine were descriptive or historical in nature (as opposed to research studies). Of the research studies identified in this preliminary survey, the large majority examined botanical therapies. Consulting additional sources of information on Ayurveda, we also examined texts in the field of Ayurveda, both lay and professional.

The consensus of the experts we consulted and the results of the preliminary work outlined above suggested that the most useful and effective way to maximize the yield of Ayurvedic studies from a literature survey was to focus on herbs. Our preliminary work also identified which disease states were most common in the Western literature on Ayurvedic therapy. This showed that these diseases are described in Western diagnostic terminology and not in terms unique to Ayurveda. The results of the PubMed search are summarized in Table 6.

Table 6. Results of preliminary search in PubMed.


Table 6. Results of preliminary search in PubMed.

Table 7 presents the disease states described by the MeSH3 headings in the 35 articles identified as clinical trials in Ayurveda from the PubMed search. Diabetes, liver disease, and ocular problems were the conditions with the largest number of articles.

Table 7. Preliminary search, clinical trials by disease state.


Table 7. Preliminary search, clinical trials by disease state.

On the basis of this preliminary nonsystematic search, we decided that a more comprehensive and systematic survey of the Ayurvedic literature was justified. We then moved forward with our initial systematic search.

Initial Systematic Search of the Ayurvedic Literature

Guided by the results of the preliminary search outlined above and with input from our panel of technical advisors, we used the search term Ayurveda plus the names of 16 major botanicals characteristically used in Ayurveda for the initial systematic search of the Western literature. We added terms for 16 herbs that we identified in published articles, including ones recognized as important by experts. The herbal terms were added to the search to increase its sensitivity, making it possible to find studies that used Ayurvedic herbal therapy without necessarily being directly identified as Ayurvedic studies.

The search was done hierarchically and initially looked for articles published in English. The focus was Ayurvedic medicine for common Western medical diseases. The search terms were Ayurveda or Ayurvedic or the scientific names of the most common botanicals: Adhatoda vasica or Albizzia lebbeck or Andrographis paniculata or Bacopa monniera or Coleus forskohlii or Commiphora mukul or Crataeva nurvala or Gymnema sylvestre or Hemidesmus indicus or Inula racemosa or Phyllanthus amarus or Picrorrhiza kurroa or Terminalia arjuna or Tylophora indica or Withania somnifera.

Table 8 shows the databases we used, and the time periods covered, for the initial systematic search of the Ayurvedic literature. The exact search terms used are listed in Appendix C.

Table 8. Database search characteristics.


Table 8. Database search characteristics.

Ayurvedic Review Strategy

Initial Screening

We designed a screening form for the initial systematic search that details characteristics we intended to extract from the articles. This form includes questions about data source (whether the information was gathered from the article's abstract or from the article itself); subject of the article (to screen out studies that were clearly not on Ayurveda); language (English, European language, Indian language); focus (whether the article specifically attempted to study Ayurvedic modalities or used an Ayurvedic therapy in the course of studying a disease state or body system); body system(s) or disease states studied; outcomes measured; Ayurvedic modalities used; subject population; and study design. A copy of the screening form is found in Appendix D.This initial review of the literature considered only articles for which there was an abstract. Two reviewers independently completed the screening form and together compared their answers, reconciling disagreements by consensus.

When screening of the abstracts was reasonably complete, we analyzed the data to describe the general characteristics of the Ayurvedic field. This was an important first step in defining our focused review.

Results of the Initial Systematic Search

We downloaded 2,565 citations from the online searches into a Microsoft Access database. Of these citations, abstracts were available for 1,562, and the remainder (1,003 citations) were obtained as titles. We initially only screened those articles for which we had abstracts. Thus, in a sense, the lack of an abstract was an exclusion criterion for the initial review.4 Of the 1,562 articles for which we had abstracts, 1,214 were related to Ayurvedic medicine and met the inclusion criteria. Each of these abstracts was independently evaluated by two reviewers using the screening form.

Studies were excluded from further review if they were:

  • Treatment studies having to do with veterinary treatment of animal disease.
  • Botanical studies having to do with the growth or botanical identity of a plant.
  • Studies in which the plant was used as a pesticide or fungicide.
  • Studies not using Ayurvedic techniques, philosophies, or materials.
  • Articles in which Ayurveda was not the main or a major therapeutic focus (i.e., survey articles of complementary and alternative medicine in general).

See Diagram 1 at the end of this chapter for a summary of the steps taken in the initial systematic search.

Diagram 1. Ayurveda Literature Search and Review Strategy: Initial Systematic Search.


Diagram 1. Ayurveda Literature Search and Review Strategy: Initial Systematic Search.

We classified 47 percent of the 1,214 abstracts as pharmacological studies. Clinical studies represented 33 percent of our abstracts; and of those, 63 percent were thought to represent controlled trials. The study designs were broken down into the categories shown in Table 9.

Table 9. Study designs identified in initial systematic search.


Table 9. Study designs identified in initial systematic search.

The articles identified as clinical studies were of special interest, since these were the types of studies we would focus on for an in-depth review. We identified the following clinical study types in our initial abstract screening:

  • Controlled trials—These include randomized controlled trials (RCTs), where subjects are randomly assigned to an intervention. We were particularly interested in RCTs. However, given their probable scarcity, we broadened this design category to include controlled clinical trials (CCTs), where the allocation is not done randomly, as well as any study that might have a comparison arm. Examples might be a study that compares patients treated with an Ayurvedic therapy to patients treated in some other way or a study that has two groups treated with different therapies.
  • Case control study—This is a study in which subjects are chosen because they have a disease, and they are compared to a comparison arm that does not. None of the studies we found fell into this category.
  • Case series—This category includes both series and single case studies without a comparison arm. Such studies involve a simple series of sequential cases (or a single case). This category accounted for 11 percent of the clinical studies. Distinctions between descriptive and experimental case series cannot be made at this stage.
  • Unclear clinical design—This category includes those studies where the two reviewers were unable to determine the study design based on the abstract alone.

Table 10 shows the distribution of target body systems or conditions identified in the 1,214 articles. One column shows the distribution for all articles, and the other lists the distribution for the 247 potential controlled trials that are of primary interest to this report. This breakdown was used to identify the potential focus areas for our comprehensive review. A single article can focus on more than one body system or condition; therefore, it is not useful to total the columns in Table 10.

Table 10. Body systems or conditions identified in initial systematic search.


Table 10. Body systems or conditions identified in initial systematic search.

The most common body systems or conditions we identified in our Ayurveda literature search were diabetes, liver/hepatitis, infectious disease, and cardiovascular disease; there were 120, 112, 103, and 63 articles, respectively. Diabetes and hepatitis had the largest number of potential controlled trials, 44 and 42, respectively, followed by central nervous system disorders and hypercholesterolemia with 23 articles each.

Table 11 identifies the various Ayurvedic modalities and therapies represented in the literature. We show the distribution of modalities for all articles and also for those articles containing potential controlled trials. Botanical modalities were the most common, accounting for 88 percent of all the studies and 95 percent of the possible comparative studies that were reviewed. This is not unexpected since the search was constructed using specific botanical names to increase the yield of Ayurvedic articles. No other modalities had significant numbers of articles, including some that are considered important in treatment of Ayurvedic patients, such as yoga or panchakarma. Some articles may have reported the use of more than one modality.

Table 11. Ayurvedic modalities identified in initial systematic search.


Table 11. Ayurvedic modalities identified in initial systematic search.

Based on the results of our detailed analysis of the data extracted from the 1,214 abstracts, and in consultation with the funding agencies, we chose diabetes as the focus of our comprehensive review. This decision was based on the fact that diabetes was the focus of the largest number of articles and had the most potential controlled trials. In addition, it has well-established diagnostic criteria and measurable outcomes that could be extracted from the Ayurvedic clinical studies. Further, diabetes is well known in Ayurvedic medicine, and the concordance of the Ayurvedic diagnosis with the Western diagnosis was close (see “Ayurvedic Diagnosis and Treatment of Diabetes” in Chapter 1).

Search of the Indian Literature

Several of the experts we consulted were of the opinion that a large body of literature existed in India that could not be obtained through the conventional search strategy outlined above. Hence, we decided after consultation with the sponsoring agencies to conduct a search for English-language Ayurvedic literature from the Indian subcontinent in December 1999-January 2000. Our goal was to ascertain how much research literature exists in Indian publications and at Indian institutions, and whether it is readily available and can be obtained in a cost-effective manner. An SCEPC staff member, a physician fluent in English and Hindi who had trained in India, went to India with the objectives described below.


  1. To establish contacts with any institutions/individuals who may be helpful in obtaining literature.
  2. To identify the major institutions involved in Ayurvedic research in India.
  3. To document the extent of the literature available in India that was not available in the databases available in the United States.
  4. To document the various languages in which this literature is available .
  5. To describe the accessibility of this literature and the difficulties in accessing it
  6. To obtain a list of journals that publish articles on Ayurveda.
  7. To get sample copies of journal publications and other, more non-traditional sources of information available in India (e.g., private communications, conference proceedings, thesis papers, etc.) to assess availability and quality.

Concurrently with this search in India, the holdings of the National Library of Medicine were also reviewed for available Indian material. The holdings turned out to be limited.


Objective 1

The following types of experts were identified and interviewed regarding our objectives:

  • Librarians of biomedical libraries
  • Researchers in Ayurvedic medical schools
  • Librarians of Ayurvedic medical schools
  • Administrators of Ayurvedic research institutes
  • Ayurvedic clinicians
  • Pharmaceutical companies manufacturing Ayurvedic herbal products

Seventeen structured interviews were conducted with the individuals listed in Appendix E. The interview questions are listed in Appendix F. An agent was identified in New Delhi who assisted in arranging some of the interviews and agreed to acquire materials in India at our direction and send them to the United States.

Objective 2

Visits were made to major Ayurvedic facilities in Mumbai, Hyderabad, Jamnagar, and New Delhi. The Indian experts identified the Central Council for Research in Ayurveda and Siddha (CCRAS) in New Delhi as an important center. They also identified Banaras Hindu University as another leading center for Ayurvedic research, but time constraints did not permit a site visit.

Objectives 3 and 4

A significant amount of literature on Ayurveda exists in India that is not available in Western databases, based on commentary from our experts in the United States and in India, as well as on the materials identified in India. This material is published in the following formats:

  • Articles in journals not indexed in Western databases
  • Thesis work done by postgraduate Ayurveda students
  • Seminar proceedings
  • Written personal communications from experts

Our reviewer examined representative samples from each category listed above. He found journals, not indexed in the West, that contained studies of sufficient quality to be included in a systematic review of Ayurvedic therapy. An important book, Researches in Ayurveda: A Classified Directory of All India P. G. and Ph.D. Theses of Ayurveda by Baghel (1997), was also identified. This work lists all postgraduate theses in Ayurveda from 1908 to 1997. A random sample of selected graduate medical theses from Gujarat Ayurvedic University in Jamnagar was screened for quality, but generally these theses were of short duration, poor design, and poorly presented; thus they were not included in this review. Written personal communications and seminar proceedings were generally very brief and not complete enough to provide useful data for a systematic review.

After viewing the holdings at a number of Indian libraries and consulting with the Indian experts, our reviewer concluded that, although literature is published in several Indian languages, the type of literature most appropriate for inclusion in our analysis (i.e., studies conducted using the Western scientific method) are generally published in English. According to the experts we consulted, the literature in Indian languages generally comprised discussions of Sanskrit texts or historical reviews.

Objective 5

The infrastructure needed to access the Indian literature is not well developed. There are no computerized databases of information equivalent to the ones that exist in the West. The only centralized listing of Ayurvedic literature identified was an abstraction service available at the CCRAS library. The librarian compiles the abstracts by hand. Articles are identified by using the holdings of the library and by recording any additional material sent to CCRAS. These reports have been published quarterly for approximately the last 10 years.

Objectives 6 and 7

The journals that comprise the CCRAS core collection are listed in Appendix G. A list of all materials eventually received from India is provided in Appendix H.

Conclusions From the Indian Literature Search

  • Literature is available in India, and some is of sufficient quality to merit inclusion in a systematic review.
  • Access is difficult because of the lack of a developed infrastructure. Using the abstracts compiled at the CCRAS library seems to be the most efficient way to identify potentially appropriate Indian studies. This view was held by both our reviewer and the Indian experts we consulted.
  • Limiting the language to English would decrease the literature yield, but we did not see evidence that it would significantly decrease the availability of studies most likely to be included in our review. On the other hand, it would greatly increase the ease of analysis.

Based on these conclusions, and with the consent of the sponsoring agencies, we obtained the CCRAS abstracts for the last 20 years and included this material in our focused search.

Focused Search of the Ayurvedic Literature

After choosing diabetes as our focus topic, we conducted a focused literature search to identify articles we might have missed in our initial systematic search and Indian literature search.


We relied on several sources for this focused analysis of diabetes in the Ayurvedic literature:

  • The initial systematic search, which already had identified 120 articles on diabetes and Ayurveda.
  • The in-person literature search in India—We obtained 16 volumes of abstracts from various Indian Ayurvedic journals, as compiled by the CCRAS library. Again, two independent reviewers evaluated all the abstracts for relevancy to the Ayurvedic treatment of diabetes. This screening yielded 318 titles. Twenty-eight articles met our inclusion criteria for human clinical trials of Ayurvedic therapy for diabetes.
  • An additional online search for common botanicals used in the treatment of diabetes—From texts of Ayurvedic herbal medicinal plants and from our early reviews of the Ayurvedic diabetic studies, we identified a list of herbs that were most often used to treat diabetes. We ran a new search focusing on herbs that are commonly used to treat diabetes. This search is also described in Appendix C. When the search was limited to human subjects only, we identified 773 additional articles.
  • A search of the online CINAHL® database using our original search terms—This search yielded 70 articles.
  • The reference lists of important articles in the Ayurveda/diabetes literature—We identified review articles of herbal treatment of diabetes and checked to see if they focused on any of the Ayurvedic herbal therapies we had identified for diabetes. Checking the bibliographies of all the articles we identified from any source, we found an additional 30 titles that were potentially relevant to our search.

As a result of our focused search, we considered a total of 1,311 articles. Two reviewers independently screened each title for relevance to Ayurvedic treatment of diabetes, and disagreements were resolved by consensus. We accepted 73 citations as either definitely or possibly relevant to the topic of our focused review; the remaining were excluded as clearly irrelevant.

All 73 accepted articles were screened using the quality review form shown in Appendix I.

Reasons for Exclusion

Of the 73 reviewed articles, 19 were deemed inappropriate for inclusion in the final data set for analysis. The reasons for exclusion were:

  • Articles containing only information about Ayurveda or treatment of diabetes, instead of original research, were considered background articles or review articles and rejected for further analysis.
  • Articles describing pharmacological, animal, and in-vitro studies were excluded.
  • Duplicate studies (those entered twice in the database) were excluded, as well as studies that appeared in different journals but contained the same or similar data.
  • Articles that had no relevant outcomes, were unobtainable, or were case reports of a single patient, were excluded from subsequent analysis.

See Diagram 2 at the end of this chapter for a description of the filtering strategy used to identify articles focusing on the Ayurvedic treatment of human diabetes.

Diagram 2. Ayurveda Literature Review Strategy of Focused Search for Human Diabetes Articles.


Diagram 2. Ayurveda Literature Review Strategy of Focused Search for Human Diabetes Articles.

Data Extraction and Synthesis

The information from the quality review form (Appendix I) and the supplemental review form (Appendix J) provided the data for the analysis that follows.

The majority of the studies that we identified were neither RCTs nor CCTs, and they were highly heterogeneous in terms of type of diabetes and treatments studied. Therefore, it was not appropriate to perform a meta-analysis. Consequently, our in-depth analysis of these studies is, by necessity, qualitative.

The qualitative analysis involved review of each article in terms of the condition studied; the diagnostic criteria used to make the diagnosis (Ayurvedic or Western or both); the location in which the study was done; the subject population (age, gender, demographics); the sample size in each arm of the study; the interventions used (primarily the names of the individual herbs or mineral and the manner of their preparation); the length of the study; and the outcomes.

All human studies that utilized Ayurvedic therapy or herbs for the treatment of diabetes were included in the Evidence Tables and in the synthesis of the evidence reported in Chapter 3, “Results.”

Selection of Studies for Further Analysis

We felt that further analysis could be done on a number of studies to allow a more direct comparison of the effects of different therapies. From the studies that had a complete review, we selected ones for further analysis if they met the following criteria:

  • The study had to be an RCT or a CCT with a comparison arm that did not receive an herb. Where there was more than one comparison arm, a diabetic comparison was preferred. RCTs of any size were included. For CCTs, the study had to have at least one treatment arm that contained at least 10 patients.
  • The study (generally case series or cohorts) had to have arms with at least 10 patients, if it was not an RCT or a CCT.
  • The study had to test a single agent, a formula acting as a single agent, or a limited combination of products (no more than three) acting as a single agent. The agent had to be dispensed more than one time to the study patients.
  • The study had to report on at least one of three outcome measures at 30 days minimum following the start of the study: hemoglobin A1c (glycosylated hemoglobin), fasting blood glucose, or post-prandial blood glucose at either 2 hours (preferred) or 1 hour (acceptable).
  • The study had to provide sample sizes, pre- and post-intervention means, and standard deviations or standard errors separately for each arm that we included in our analysis.

For studies that reported on both 1-hour and 2-hour post-prandial blood glucose levels, we chose the 2-hour measurement for the analysis. For studies that reported more than one followup time, we selected the followup time that was closest to 3 months.

Twenty-two studies were selected for further analysis based on the criteria described above. These studies represented either RCTs (2) or CCTs (5) with a comparison arm that did not receive an herb, or studies that had pre/post5 comparison data available (15). The study types selected for further analysis are summarized in Table 12

Table 12. Summary of studies included for further analysis.


Table 12. Summary of studies included for further analysis.

Given the clinical heterogeneity of the interventions and the populations across the studies, we decided not to pool data across studies. Instead, separately for each of the three clinically relevant outcome measures (fasting blood glucose, post-prandial blood glucose, and hemoglobin A1c), we estimated a common effect size and its confidence interval for each study.

We considered two subgroups of eligible studies separately: the RCTs and CCTs that had nonherbal comparison arm(s) were considered one subgroup, and all other eligible (pre/post data available) studies made up the second subgroup. The RCT/CCT subgroup allowed us to do a comparison between herbal and nonherbal arms within a study, providing a stronger level of evidence than pre/post comparisons. For these studies, we were able to compare the effects of different treatments after the placebo or nonherbal arm results were taken into account. The pre/post studies subgroup did not allow a concurrent comparison, but the difference between pre- and post-treatment values could be estimated.

The results of this analysis are displayed graphically for ease of comparison across studies and are presented in Chapter 3. Appendix K summarizes the data and methods used to calculate the common effect sizes.


MeSH stands for Medical Subject Headings, a controlled vocabulary derived by the National Library of Medicine used for searching major medical databases.

However, all articles without abstracts were considered for inclusion after we selected our focus topic (diabetes).

The term “pre/post data available” refers to studies that measured one or more of the three outcomes of interest (fasting blood glucose, post-prandial blood glucose, and/or hemoglobin A1c) both pre-intervention and post-intervention. This subgroup included CCTs, cohorts, and case series.


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...