Tibetan Medicine: A Systematic Review of the Clinical Research Available in the West

Background. Little is known about Tibetan medicine (TM), in Western industrialized countries. Objectives. To provide a systematic review of the clinical studies on TM available in the West. Data Sources. Seven literature databases, published literature lists, citation tracking, and contacts to experts and institutions. Study Eligibility Criteria. Studies in English, German, French, or Spanish presenting clinical trial results. Participants. All patients of the included studies. Interventions. Tibetan medicine treatment. Study Appraisal and Synthesis Methods. Included studies were described quantitatively; their quality was assessed with the DIMDI HTA checklist; for RCTs the Jadad score was used. Results. 40 studies from 39 publications were included. They were very heterogeneous regarding study type and size, treated conditions, treatments, measured outcomes, and quality. Limitations. No Russian, Tibetan, or Chinese publications were included. Possible publication bias. Conclusions. The number of clinical trials on TM available in the West is small; methods and results are heterogeneous. Implications of Key Findings. Higher quality larger trials are needed, as is a general overview of traditional usage to inform future clinical trials. Systematic Review Registration Number. None.


Background
Traditional Tibetan medicine (TM), sometimes called "Lamaist" or "Buddhist" medicine, has developed in 1200 years into a unique medical system [1][2][3]. In TM, disease is understood as an imbalance of the three "Nyes-pa" (principles) consisting of one or two elements: "rLung" (air, wind), "mKhris-pa" (fire), and "Bad kann" (earth and water) [4]. Buddhist philosophy as well as shamanic origins of Tibetan culture form a background of cosmological, mind-body, and spiritual dimensions [1][2][3]. Treatment may consist of medicines (usually preparations of plants [5], seldom minerals or animals), physical treatments (e.g., massage, baths), life and diet regulation, or spiritual techniques [4]. Standardization of the originally individualized medicines, separation from the underlying philosophies, and discontinuation of some techniques (e.g., Tibetan dental medicine, cauterization) have led to derivative forms of TM [6]. We will use the term "Tibetan medicine" for the traditional TM (with its individual life style advice, diet, physical, and spiritual means) as well as larger or smaller subsets or varieties of it, down to single formulas.
Besides the regions of the historical Tibet, very similar medical traditions are practised since the Mongolian conquest of Tibet in the 13th century in Mongolia, adjacent Siberia, and in the Russian province Kalmykia (Figure 1) [7]. Especially with traditional Mongolian medicine, TM has a substantial similarity. TM use in Western industrialized countries (the "West") originates in a line of descendants of a Buryat physician migrating westward in the 19th century ( Figure 1) [8,9]. Still, there is little awareness of TM in the general Western public. Following the rising interest in traditional Chinese medicine (TCM) and complementary or alternative medicine (CAM) in general, more demand from Western countries can be expected in the future. The amount of available research in the West is small. A Medline search up to December 31, 2010, for example, for "Tibetan medicine" returned 371 hits, 0.0183 times the number for "traditional Chinese medicine. " The existing literature indicates a palliative, possibly curative potential, especially for chronic diseases [10], but studies on its multimodal individualized approach are scarce and systematic reviews exist only for one TM product [11][12][13][14][15]. Therefore, we attempted to present in this paper a systematic overview of clinical research currently  I   II   III   III III (1) (2)   I   II   III   I I I I I I III I I I I I I I I I I I I I I  III I I I I I I I I I (1)  available in the West on Tibetan medicine, and aim to provide details on methods and study quality. Some preliminary data can be found in [16].

Methods
A preliminary list of 15 literature databases was tested using the search terms "Tibetan medicine, " "Himalaya medicine, " "Tibetan herbal, " and "Lamaistic medicine. " The database list had been compiled from recommendations by experts, by Ovid [23], and by Deutsches Institut für Medizinische Dokumentation und Information (DIMDI) [24]. Those returning the most hits were used for the literature search, together with databases that were recommended by experts or appeared relevant in their characterization on the websites of DIMDI or the Charité library [25]. We finally searched seven databases up to publication date December 31, 2010: ABIM (accessed via Rijksuniversiteit Groningen), AMED (DIMDI), CAMbase (cambase), CCmed (DIMDI), Cochrane Collaborative Library (OVID), Embase (OVID), and Medline (PubMed). The search term "(Tibet OR Himalaya OR Mongolia OR Buddhist) AND (herbal OR medicine) AND study" was adapted as necessary to database language and syntax. Similar searches were used on the medical information services of DIMDI [24] and ZB MED [26] and by adding "AND clinical study" on Google scholar [27]. The published literature lists [28,29]  Included studies = 40 primary outcome was defined, the first outcome mentioned in the title or the abstract was extracted, unclear cases were discussed with another author (C. M. Witt) until consensus was reached.
Methodological quality of the studies was determined with a DIMDI checklist ( Table 1) that is used to evaluate studies for in-/exclusion in health technology assessments (HTA) in Germany [35]. The checklist has up to 31 items sorted into 7 categories and was used on a descriptive basis. Randomized controlled trials (RCTs) were further evaluated with the Jadad score [36,37]. Descriptive statistics were calculated using MS Access 2003 andMS Excel 2003 [33, 34].

Results
From 1383 screened records, we identified 40 studies reported in 39 publications (one contains 2 studies [38]), see Figure 2. An additional search without the terms "herbal, " "Buddhist, " and "Mongolian" did not result in fewer relevant publications. Thirty-five of the publications were journal articles, two were book chapters, and one is treated in this paper as a single Internet publication, although different findings had been published in several online media reports [39]. Only 18 publications were found by the initial data base searches. Most of the others were indeed indexed, as a reverse search (for already known publications) revealed. Written in English were 53.8% ( = 21) of the publications, the other 46.2% (18) were in German. Most publications came from Poland and Switzerland (30.8% or = 12 each, all on products of Padma AG). The Asian studies were from India (15.4%, = 6) or China (5.1%, = 2). The earliest publication appeared in 1970. Since 1990 every 5 years about 3 new RCTs were published and, less evenly distributed, most of the observational studies (total = 14). The 5 nonrandomized controlled trials were published between 1986 and 1991, and the 6 case studies or case series in 1998 or later ( Table 2). The setting of 7 studies (17.5%) was multicentred [40][41][42][43][44][45][46]. Four studies (10.0%) were retrospective [40,45,47,48].
From those publications including herbal medicines, four did not provide details on the used medication [42,58,59,62], two provided the name of the preparations but not the ingredients [43,48], and two provided the name of the preparation and ingredients, but no information on the quantity of the ingredients [44,63]. Data on both ingredients and their quantity was only available for Padma 28 and Padma Lax.
The duration of the non-randomized controlled trials was between 6 weeks and 6 months (mean = 43 d), 54% of the 678 patients received the verum Padma 28. Four non-randomized controlled trials included children with chronic respiratory tract infections [46,64,65] or juvenile arthritis [66]. One trial on adults included angina pectoris patients [61].
The checklist results for quality assessment are presented at item level in Table 3 for each study. Depending on study 4 Evidence-Based Complementary and Alternative Medicine (A) Selection of participants Participants (1) Were the criteria for in-/exclusion defined sufficiently and clearly? A1 (in-/exclusion) (2) Were the criteria for in-/exclusion defined before intervention? A2 (predefined) (3) Was the health status recorded in a valid and reliable way?
A3 (health status) (4) Were the diagnostic criteria of the disease described?
A4 (diagnostic criteria) (5) Were the studied/exposed patients representative for the majority of the exposed population or the "standard users" of the intervention? A5 (representativity) (B) Allocation and study participation Allocation (1) Were the exposed/cases and nonexposed/controls from the same base population? B1 (basic population) (2) Were intervention/exposed and control/nonexposed groups comparable at baseline? B2 (comparable) (3) Was allocation randomized, with a standardized procedure? B3 (randomization) (3) Was if assured that participants did not crossover between intervention and control group? D3 (no crossover) (E) Outcome recording Outcome (1) Were patient-centred outcome parameters used? E1 (patient-centred) (2) Were the outcomes recorded in a valid and reliable way? E2 (recording) (3) Was outcome recording blinded? E3 (blinded outcomes) (4) For case series: was the distribution of prognostic factors recorded sufficiently? E4 (prognostic factors) (F) Drop-outs Drop-outs (1) Was the response rate in intervention/control group sufficient, or, for cohort studies, could a sufficient part of the cohort be tracked for the full study duration? F1 (evaluable number) (2) Were the reasons for the dropouts of participants stated? F2 (reasons) (3) Were the outcomes of dropouts described and included in the analysis? F3 (outcomes) (4) If differences were found: were they significant? F4 (significance) (5) If differences were found: were they relevant? F5 (relevance) (G) Statistical analysis Statistics (1) Were the described analytic methods correct and the information sufficient for a flawless analysis? G1 (correct) (2) Were confidence intervals given for means and for significance tests? G2 (CIs given) (3) Were the results presented in graphical form, and were the underlying values stated? G3 (graphics) * Translated from [35], * * used in Table 3. type and setting, 10 to 26 items could be answered. Had the assessment been for HTA purposes, only 1 case study [76] and 1 RCT [55] would have been eligible for inclusion in a HTA. Ignoring only one item (G2, provision of confidence intervals) would have raised that number to 13, including 8 RCTs that the Jadad score rated as good or very good quality. The Jadad score of the 15 RCTs (Table 4)        Randomized controlled    All studies followed conventional "Western" medical diagnoses. Additional traditional TM diagnostics were recorded in 11 studies that investigated the traditional multimodal treatment. In 9 of them, the Tibetan diagnosis was used to plan the therapy [39,42,43,45,48,58,59,62,76].
Assessed outcomes included clinical outcomes such as symptom scales ( = 37 studies), laboratory tests (19), clinical tests (such as ankle/brachial pressure index, blood pressure, or weight; 9), and other (9), such as microbiology, histology, or the need for conventional medication. The authors drew positive conclusions on their data in 34 studies. In 2 RCTs, TM was found to be inferior to conventional medicine, but better than placebo [44,46]. In one study, only 1 of 5 outcomes improved [60]; in 2 studies the primary outcome did not change significantly while secondary outcomes did [42,52]. The comparison of the traditional and a not further specified "special" Tibetan medicine [59] resulted in comparable clinical improvements. The remaining studies found no significant differences to controls [49,65], or their authors were doubtful about the observed effects [39]. Statements about adverse effects were included in 23 studies, in 11 of them no adverse effects were reported, and 2 studies did not mention the number of patients with adverse effects [39,53]. The remaining 10 studies reported adverse effects with a range from 5% to 55% of the patients.
Osteoarthritis or rheumatoid arthritis was treated in three trials: 1 RCT [58] and 1 observational study [62] with the traditional multimodal approach, and with Padma 28 in 1 controlled trial [66]. All studies reported pre-/postimprovements or superiority to controls regarding symptom severity.
In 3 other trials, hepatitis B patients were either treated with a "special" TM (that was not further specified) in comparison to traditional TM (1 RCT [59]) or with Padma 28 (2 observational studies [47,69]). All publications reported positive results for laboratory outcomes. The comparison of traditional and "special" traditional TM found comparable improvements but did not achieve seroconversions.

Discussion
In this paper, we presented an overview of the clinical research on traditional Tibetan medicine (TM) that is currently available in the West. Three quarters of the included studies tested single formulations, most of them products of Evidence-Based Complementary and Alternative Medicine 13 a single company. One quarter investigated the traditional multimodal TM approach. Studies were very heterogeneous regarding study type and size, treated conditions, treatments, measured outcomes, and quality.
In this, to our knowledge, first systematic overview of clinical TM research available in the West, we tried to minimize subjectivity using pre-defined systematic methods wherever possible (data extraction sheets, established quality assessment tools). However, the small number of trials scattered over a whole medical system and very heterogeneous treated diseases prohibited more formal or in-depth analyses.
Despite the broad literature search, some studies may not have been identified, for various reasons. Although Mongolian and Tibetan medicine are not completely identical, we have included "mongolian" in the search terms in order to find as much relevant literature as possible. We did not search for single TM interventions such as bathing or bloodletting and assumed that they are well covered under the umbrella term "medicine. " Although we detected with this search a study on Tibetan yoga [60], we possibly missed other studies. Furthermore, publication bias could have had occurred, as some papers [11,15,58] indicated the existence of studies that have not been published (or at least not in indexed journals) [77][78][79][80][81][82]. Several papers were not identified by our search strategy in the literature databases, but could have been found searching for " Padma 28" or "Padma Lax. " Clearer labelling of TM studies in the future would be helpful. On the other hand, our search seems to have been partly redundant, as all identified publications could have been found with fewer search terms. The main limitation is that our language restriction excluded articles in Russian, Tibetan, and Chinese. This literature was not accessible for us. Furthermore, we learned from our field work and from discussions with Western and Chinese manufacturers during an interdisciplinary symposium on TM [16] that most literature on clinical research published in Tibetan is not available in indexed journals and that most research published in Chinese addresses preclinical questions.
The evaluated literature presented a high number of studies without a control group. Only a few single products were subject to in-depth investigation. Both facts indicate an early stage of research in a new and largely unexplored field where only few focused inquiries exist. The predominating countries of origin (>2/3 European) and the 70% of studies on Padma products among the included literature are consequences of the language restrictions of our search as well as of the historical development of TM utilization in the West. Although they are prescribed in a standardized and nonindividualized fashion, the Padma products are a genuine Tibetan medication according to manufacturers, study authors, and independent experts [17,83,84]. Adaptation of constituents to local situation and ecology is an accepted practice in TM. It was done in one study when Tibetan physicians reduced the traditional Byu-Dmar 25 by 12 ingredients to comply with Tibetan pharmacopoeia and European regulations, resulting in Byu-Dmar 13 [63]. A similar strategy might have been used in two other studies [39,62].
The heterogeneous nature of the included studies demanded the use of quality assessment instruments that were suitable for diverse study designs, but have the general disadvantage of allowing only rough estimates of the assessed quality. Nevertheless, they allowed spotting the more obvious deficiencies that are symptomatic of research at an early stage and that future research can avoid with improved methodology on the grounds of evidence-based medicine. Case studies and observational studies are useful to gather information on traditional usage and settings and to identify areas where controlled studies seem promising. Then, to provide higher-level evidence, more RCTs will be needed. Methodological issues such as small samples, insufficiently described populations in many studies, pre-/post-comparisons of treatment within a group, or comparator treatments without clinical relevance all indicate that TM research as seen through the Western literature is still at a nascent stage. Furthermore, the quality of most studies and the heterogeneity of interventions and outcomes make clear conclusions impossible.

Conclusion
The clinical research on traditional Tibetan medicine (TM) that is available in Western industrialized countries is scarce and scattered over a whole medical system, but shows interesting results. Better research methodology should be applied, and larger trials are needed, as is a general overview of traditional usage to inform future clinical research.