Ayurveda

Textual foundations

Historians cited by Ayurveda faculty at Banaras Hindu University generally place the documentation of the core Ayurvedic literature between about 400 BCE and 200 CE, while accepting that its theoretical framework has older roots. A separate overview by a former professor at the College of Indian Medicine, Madras, describes a Samhita period from the 6th century BCE to the 7th century CE and attributes the origin of Ayurveda to the Atharva Veda, which is a tradition-internal attribution and not a documented authorship. Both are secondary summaries and neither dates individual works.

Two compendia are treated as classical foundations. A 2016 review states that Agnivesha compiled knowledge from the Vedas and that Charaka and other scholars edited it into the Charaka Samhita, and that the Sushruta Samhita describes surgery. The sources opened for this draft give no firm dates for the individual authors, so the authorship and dating of both works remain unsettled here.

Manuscript evidence

The surviving physical record is later than the composition range above. A transcription page on the Suśruta Project's wiki records a palm-leaf manuscript of the Suśrutasaṃhitā in Nepalese script, held in the Kaiser Library, Kathmandu (KL 699), as dated Saṃvat 301, which the page equates to 878 CE. The page does not discuss when the text itself was composed, and scholarly datings of the Charaka and Suśruta compendia were not checked against specialist histories of Sanskrit medicine.

Institutional setting in India

A 2026 review protocol by authors at Indian government research bodies describes Ayurveda as one of the Ayush systems, alongside yoga and naturopathy, Unani, Siddha, Sowa Rigpa and homeopathy, collectively governed under the Ministry of Ayush. The World Health Organization published a 61-page document on benchmarks for training in Ayurveda in 2010; its contents were not opened for this draft. Official Indian curriculum pages could not be opened either, so national training standards are not described here.

The name and the doshas

A 2013 review of lead-based preparations glosses Ayurveda as 'knowledge of life'. In research papers the word is used both for the whole system, with its own diagnosis, regimens and procedures, and for individual products labelled as Ayurvedic, so a finding about one product or procedure does not describe the system.

Within the tradition, three doshas, vata, pitta and kapha, are central explanatory categories. A 2016 review glosses vata as maintaining cellular transport, electrolyte balance and elimination of waste, pitta as regulating body temperature and hunger and thirst, and kapha as lubricating the joints. These are the authors' translations of tradition-internal concepts into physiological language, and they are not measured biomedical constructs.

Related terms

Agni, usually rendered as digestive fire, appears in a 2023 history-of-physiology paper discussing digestion and metabolism, and a 2025 stress review links disturbed agni to ama, described there as toxin accumulation. Both papers are written by Ayurveda-trained authors and present these concepts as the tradition's own explanations.

Rasa shastra is described as the practice of combining herbs with metals, minerals and gems. Bhasma denotes a prepared ash, for example Naga bhasma, described as lead sulphide ash. 'Herbo-mineral' and 'herbo-metallic' are descriptive labels used in the research literature for such products.

Assessment and individualisation

Ayurvedic care is described as individualised. In a reliability study nested in a knee osteoarthritis trial, four Ayurvedic experts each took a semistructured history from 30 patients and reached consensus diagnoses covering disease-related and constitutional entities. A rheumatoid arthritis pilot trial also gave individualised Ayurvedic therapy and developed a placebo for each Ayurvedic dosage form. These are research settings and do not describe every clinic.

Clinic procedures

A 2016 review lists five Panchakarma actions as virechana (purgation), vamana (therapeutic emesis), basti (enemas of medicated oils), rakta moksha (described as blood detoxification) and nasya (nasal medication). Other descriptions may group the procedures differently.

In a small uncontrolled report of ten people with cerebellar ataxia, treatment combined shirodhara or shirobasti with a medicated oil for 45 minutes daily, followed by abhyanga with the same oil and steam, for 14 days, plus oral medicines for a month. The package cannot show what any single component contributes. In a German trial, Ayurvedic treatment meant 15 multimodal treatments over 12 weeks.

Daily routines and oil pulling

A 2025 review describes dinacharya as daily routines aligned with circadian rhythms, alongside dietary guidance emphasising warm, nourishing foods. Oil pulling trials used edible oils, half of them sesame, for 7 to 45 days.

The eight branches

A 2016 review lists eight disciplines, called Ashtanga Ayurveda: internal medicine, psychological disorders, paediatrics, Rasayana (described as geriatrics), aphrodisiacs, surgery, ear, nose, throat and eye care, and toxicology. The list shows the breadth of the tradition's own scope. It is not a list of conditions for which any treatment has been tested.

Uses described by authors and practitioners

A 2025 review presents stress as a disruption of equilibrium expressed through the doshas and proposes dinacharya, abhyanga, shirodhara, Panchakarma and herbs. A 2013 review notes that Naga bhasma has been used in India for diabetes, spleen enlargement, diarrhoea and skin diseases. These are descriptions of tradition and practice, and the reviews do not demonstrate benefit.

A 2025 review of Triphala, a three-fruit Rasayana formulation, describes it as used for gastrointestinal health and rejuvenation. The review's authors present these as established attributes, but the review is not itself a controlled test.

Conditions that have been studied

Randomised trials and systematic reviews exist for type 2 diabetes, rheumatoid arthritis, knee osteoarthritis, essential hypertension and oral hygiene practices. They are summarised below. Each applies to the specific products or procedures tested.

Type 2 diabetes

A 2011 Cochrane review included seven randomised trials, six of proprietary herbal mixtures and one of whole-system Ayurvedic treatment, with 354 adults with type 2 diabetes treated for 3 to 6 months. HbA1c, fasting glucose or both fell with Diabecon, Inolter and Cogent DB against placebo or no added treatment, but not with Pancreas tonic or Hyponidd. The reviewers could draw no definite conclusions because of weak methods and small samples.

A 2022 review searched 18 databases to January 2021 and found 199 randomised trials of 98 Ayurvedic medicines in 21,191 participants. Reporting of methods was often poor. HbA1c fell against control with several named plant medicines, by between 0.3 and 1.6 percentage points depending on the medicine, and the authors concluded that evidence suggests benefit for glycaemic control but is limited.

Rheumatoid arthritis and knee osteoarthritis

A 2005 review of seven randomised trials in rheumatoid arthritis, (searches to March 2003) found that of three placebo-controlled trials, one high-quality trial showed no benefit, one incompletely reported trial suggested benefit and one showed no difference. Four compared Ayurvedic medicines with each other and could not be interpreted. A 2011 double-blind pilot randomised 43 adults to methotrexate, Ayurveda or both for 36 weeks and found no significant differences on efficacy measures.

In a multicentre open-label trial in Germany, 151 adults with knee osteoarthritis received multimodal Ayurvedic treatment or multimodal conventional care for 12 weeks. WOMAC scores improved by 61.0 points (95% CI 52.4 to 69.6) with Ayurveda against 32.0 (21.4 to 42.6) with conventional care,; effects were sustained at 6 and 12 months.

Hypertension and oil pulling

A 2025 review of Ayurvedic interventions for essential hypertension included 44 studies and meta-analysed ten randomised trials with 524 participants. Against placebo, systolic pressure fell by 2.63 mmHg (95% CI -6.04 to 0.79), a non-significant difference. Against antihypertensives it was -0.22 mmHg (-0.82 to 0.38). Heterogeneity exceeded 90%, which limits the pooled figures.

A 2016 review of five small oil pulling trials found no significant difference from control in plaque or gingival scores. A 2023 meta-analysis of 25 trials in 1,184 participants reported better modified gingival scores than non-chlorhexidine mouthwashes and worse plaque control than chlorhexidine; overall certainty was very low.

Diagnosis, procedures and gaps

In the nested reliability study, experts agreed on the main diagnosis in 95% of consensus ratings, but kappa values for central diagnostic aspects ranged from 0 to 0.4, with poorer agreement on constitutional than disease-related entities. It included only 30 patients.

The sources appraised for this draft do not isolate a controlled test of Panchakarma, shirodhara, abhyanga, Rasayana or dinacharya alone. The ataxia report is uncontrolled. PubMed searches for these terms were run, but results beyond the cited papers were not appraised, so controlled studies of individual procedures may exist.

Trial design

Whole-system trials are hard to blind. The knee osteoarthritis trial was open-label, with a conventional-care comparator, and the rheumatoid arthritis work needed bespoke placebos to be blinded at all. Reviews found small samples, unreported methods and varied products, and the hypertension meta-analysis was highly heterogeneous. Each limits what can be inferred about Ayurveda as a whole.

Authorship and what remains open

Many of the historical and evidence sources here are written by authors working in Ayurveda departments or government Ayurveda bodies. That does not invalidate their findings, but the historical dating summaries come from within the field, and independent specialist history of Sanskrit medicine was not accessible for this draft.

Metals in some oral preparations

In a 2004 study, 14 of 70 Ayurvedic products made in South Asia and sold in Boston-area stores (20%; 95% CI 11% to 31%) contained lead, mercury or arsenic by X-ray fluorescence, and each could exceed regulatory intake standards if taken as directed. A 2008 study of 193 products bought online found detectable metals in 20.7%, similar for US-made (21.7%) and Indian-made (19.5%) products. Rasa shastra products had metals more often (40.6% against 17.1%), and all metal-containing products exceeded at least one daily intake standard.

Other analyses

A Dutch study sampled 292 traditional herbal preparations used in Ayurveda, traditional Chinese medicine and traditional Tibetan medicine between 2004 and 2007. Mercury, arsenic or lead was found in 186 (64%), fewer than 20% of labels suggested their presence, and 59 (20%) were estimated to give intakes well above safety limits. The samples are not Ayurveda alone. A US cluster investigation of lead poisoning found lead in 65% of 252 Ayurvedic samples, though samples collected during an investigation are not representative.

Tradition, contamination and quality control

A 2013 review notes the traditional view that elaborate preparation modifies the toxicity of a metal, while modern scientists consider such products toxic because of their metal content. It found no compelling reports of toxic manifestations from Naga bhasma, yet extremely high lead levels in elemental analysis, and said safety is difficult to assure. A 2026 review of Ayurvedic raw materials lists adulteration, species substitution and contamination as threats to quality.

Suitability and care boundaries

Adverse events were often not reported in the diabetes trials, and the 2011 review found only mild events where they were reported. None of the sources reviewed here covers drug interactions, pregnancy or children, so no statement on them is made. Anyone with a serious or worsening condition, or taking prescribed medicines, should not delay or replace necessary medical care, and may ask a clinician about any oral product. Training standards were not verified beyond the existence of the 2010 WHO benchmarks.