Self-Heal (Xia Ku Cao) in Chinese Thyroid Trials

Table of Contents

1. The Findings in Brief
2. Xia Ku Cao for Neck Swellings in Traditional Chinese Medicine
3. How the Trials Were Designed
4. Hyperthyroidism and Graves' Disease
5. Thyroid Nodules
6. Hashimoto's Thyroiditis and Subacute Thyroiditis
7. Side Effects Recorded in the Trials
8. Why the Reviewers Rate the Certainty as Low
9. Laboratory and Animal Studies
10. Iodine, Other Medicines and Safety Signs
11. What Is Not Known
12. Key Research Papers
13. Connections
14. Featured Videos

The Findings in Brief

• The human trials exist, and there are many of them. A 2025 meta-analysis pooled 17 randomised trials (1,366 patients) in hyperthyroidism; a 2026 meta-analysis pooled 9 trials (966 patients) in Graves' disease; a 2021 meta-analysis pooled 13 trials (1,468 patients) in thyroid nodules; and a 2020 Chinese-language meta-analysis pooled 11 trials (1,215 patients) in Hashimoto's thyroiditis.
• Self-heal was almost always an add-on. In these trials one group received standard thyroid medicine and the other received the same medicine plus a self-heal preparation. The trials test whether adding the herb changed the result, not whether the herb works on its own.
• The pooled results favoured the add-on group on blood thyroid hormone levels, thyroid antibodies and the measured size of the thyroid gland or nodules.
• The reviewers rate the evidence as low or very low certainty. Each review names the reasons: unclear allocation concealment, missing blinding, and trial quality that is "uncertain".
• Laboratory and animal work points the same way, in thyroid cells and in rats and mice with experimental thyroid disease. Those results describe possible mechanisms; they are not evidence of an effect in people.

Xia Ku Cao for Neck Swellings in Traditional Chinese Medicine

In traditional Chinese medicine, xia ku cao is described as bitter, acrid and cold, and it is assigned to the liver and gallbladder "meridians". A 2022 review of the plant's ethnopharmacology lists its traditional actions as clearing the liver, dispersing "fire", improving vision and dispersing swelling, with scrofula (swollen neck glands) and goitre among the classic uses. These are traditional descriptions, not findings from modern trials.

"Phlegm nodules" is a traditional Chinese medicine term for lumps that the tradition attributes to stagnant phlegm, a category that included what today would be called goitre, enlarged lymph nodes and thyroid nodules. Xia ku cao was one of the herbs repeatedly chosen to "dissipate" them, usually as one ingredient in a multi-herb formula. A 2026 review of 44 Graves' disease trials of Chinese herbal formulas found that Prunella vulgaris was the most frequently used herb in those formulas, ahead of rehmannia root, scrophularia root, fritillaria bulb, bupleurum and white peony.

Modern Chinese practice also uses single-herb preparations made from self-heal. The trial reports name two of them by their generic names: Prunella Oral Liquid (a liquid preparation described as a Chinese patent medicine prepared from Prunella vulgaris) and Prunella granules. Most of the trials on this page used one of these preparations rather than a home-made tea.

How the Trials Were Designed

Almost every trial in these reviews followed the same pattern, and understanding it is the key to reading the results.

1. Both groups received standard treatment. In hyperthyroidism and Graves' disease, that meant antithyroid drugs (the medicines that slow the thyroid's hormone production; the class includes methimazole, carbimazole and propylthiouracil). In thyroid nodules and Hashimoto's thyroiditis, it meant levothyroxine, a synthetic thyroid hormone. Some Graves' disease trials used radioactive iodine (iodine-131) instead.
2. One group also received a self-heal preparation. The comparison is "standard treatment plus self-heal" against "standard treatment alone".
3. The outcomes were mostly laboratory and ultrasound numbers. These included free triiodothyronine (free T3) and free thyroxine (free T4), the active thyroid hormones; thyroid-stimulating hormone (TSH), the pituitary signal that rises when thyroid hormone falls; thyroid antibodies; and the measured width, thickness and length of each thyroid lobe or the diameter of a nodule.
4. Several trials also used a "total efficacy rate". This is a composite judgement, common in Chinese clinical research, that combines lab results, gland size and symptoms into a single "effective / not effective" category.

The design has one important consequence. Because no group received a dummy (placebo) preparation, both patients and doctors in most trials knew who was taking the herb. The reviewers return to this point in their quality assessments, as described below.

Hyperthyroidism and Graves' Disease

Hyperthyroidism means an overactive thyroid. Graves' disease, an autoimmune condition in which an antibody called the thyrotropin receptor antibody (TRAb) switches the gland on, is its most common cause.

The 2025 meta-analysis: 17 trials, 1,366 patients

Wei and colleagues searched eight Chinese and English databases for randomised trials comparing self-heal preparations plus antithyroid drugs with antithyroid drugs alone. They assessed bias with the Cochrane tool and graded the evidence with the GRADE system (Grading of Recommendations, Assessment, Development and Evaluation), the standard scale for how much confidence a body of evidence deserves.

Compared with antithyroid drugs alone, the add-on group showed:

• lower free T3 (standardised mean difference, SMD, −0.98) and lower free T4 (SMD −0.82);
• a higher TSH (SMD 0.71), the direction expected as an overactive thyroid settles;
• a lower thyrotropin receptor antibody level (SMD −1.11);
• smaller measurements of the thyroid isthmus and of the width, thickness and length of both thyroid lobes;
• a lower inflammatory marker, tumour necrosis factor-alpha, and a higher interleukin-10, a signalling molecule that calms inflammation;
• fewer adverse events (risk ratio 0.34).

Two results were neutral: relapse rates, and the inflammatory markers interleukin-6 and interferon gamma, were comparable in the two groups. A standardised mean difference expresses an effect in units of the spread of the data, which allows trials that used different lab methods to be pooled; it does not translate directly into a hormone level a patient would see on a lab report.

The authors conclude that the combination was superior on thyroid function, antibodies, gland size, inflammation and adverse events, and in the same sentence state that "owing to the low strength of evidence from the included studies, this conclusion requires further validation in more high-quality RCTs."

The 2026 Graves' disease meta-analysis: 9 trials, 966 patients

Chen and colleagues looked only at Prunella Oral Liquid in Graves' disease, searching seven databases up to September 2025. All nine trials compared the preparation added to Western treatment with Western treatment alone.

• Added to antithyroid drugs: lower free T3 (SMD −0.94) and free T4 (SMD −0.87), higher TSH (SMD 0.79), lower TRAb (SMD −1.26) and lower thyroglobulin antibody (mean difference −105.33 IU/mL). Each thyroid lobe measured smaller by roughly 0.2 to 0.7 cm depending on the direction measured, and the isthmus by 0.14 cm. Adverse events were less common (risk ratio 0.35).
• Added to radioactive iodine: lower TRAb (mean difference −17.00 U/mL), lower thyroglobulin antibody (−132.00 U/mL) and lower thyroid peroxidase antibody (SMD −2.98).
• Added to antithyroid drugs plus nutritional supplements: lower free T3 (−0.93 pmol/L) and higher TSH (0.17 μIU/mL).

The authors write that the risk of bias came mainly from "unclear allocation concealment and performance bias due to lack of blinding", that the certainty of evidence by GRADE "ranged from low to very low", and that the findings need to be interpreted with caution until "rigorously designed multicenter RCTs with standardized protocols" are done. The published author list shows two of the eight authors with an affiliation at a pharmaceutical technology company; the paper declares no competing interests.

A network comparison and the earliest trial

A 2024 network meta-analysis by Yang and colleagues compared thirteen Chinese medicines across 35 Graves' disease trials (2,828 patients). In its ranking, Prunella granules had the largest effect on lowering free T3 and free T4, and Prunella Oral Liquid the largest effect on lowering thyroid peroxidase antibody and TRAb. A network meta-analysis ranks treatments by combining direct and indirect comparisons; its rankings carry the uncertainty of the trials underneath them.

One of the oldest indexed trials, published in Chinese in 2007, randomised 56 people with hyperthyroidism to standard treatment with or without Prunella oral solution and measured the thyroid by colour ultrasound. Both groups' glands shrank; the add-on group's shrank more, to a size no longer different from that of 20 healthy controls.

The 2025 and 2026 reviews searched overlapping databases over overlapping years. The abstracts do not say how many trials the two reviews share, and two reviews drawing on the same pool of trials are not two independent confirmations.

Thyroid Nodules

Thyroid nodules are lumps within the thyroid. Most are benign, and the trials here concern benign nodules treated with thyroid hormone tablets.

Han and colleagues pooled 13 randomised trials with 1,468 patients (databases searched to April 2021), using version 2.0 of the Cochrane risk-of-bias tool, with two reviewers working independently. Compared with levothyroxine alone, self-heal plus levothyroxine showed:

• a higher "clinical efficacy" rate (risk ratio 1.22);
• a smaller nodule diameter (mean difference −0.43);
• lower free T3 (mean difference −1.99) and lower free T4 (mean difference −3.20).

Trials that used a different thyroid hormone tablet ("thyroxin tablets") also showed a higher efficacy rate with self-heal added (risk ratio 1.29).

For side effects, the pooled risk ratio was 0.67, with a 95% confidence interval of 0.36 to 1.22. Because that interval includes 1.0, the pooled result does not show a clear difference in adverse reactions between the groups, although the abstract's conclusion describes "reducing the occurrence of adverse reactions". The authors also write that "the quality of these studies is uncertain, and higher quality and more RCTs are needed".

A 2007 Chinese trial in 98 people with goitre (56 with hyperthyroidism, 24 with hypothyroidism and 18 with simple goitre) reported that thyroid size fell in both groups, and more with Prunella Oral Liquid added to standard treatment. That report gives no blinding or side-effect details in its English abstract.

Hashimoto's Thyroiditis and Subacute Thyroiditis

Hashimoto's thyroiditis

Hashimoto's thyroiditis is an autoimmune condition in which antibodies against thyroid peroxidase (TPO antibodies) mark a slow attack on the gland, often ending in an underactive thyroid. A 2020 Chinese-language meta-analysis by Zhang and colleagues pooled 11 randomised trials with 1,215 patients that compared a xia ku cao preparation plus levothyroxine with levothyroxine alone.

• The "total efficacy" rate, which combined thyroid function, antibodies, gland and nodule size, and symptoms such as fatigue and cold intolerance, was higher with the herb added (risk ratio 1.15).
• TPO antibody levels were lower (SMD −0.91).
• The left lobe, right lobe and isthmus of the thyroid measured smaller.

The authors graded the quality of all the evidence as low or very low. They note that a trial sequential analysis showed the total number of patients had reached the size needed for the comparison, but that the pooled results "may be affected by one study with high bias risk".

Subacute thyroiditis

Subacute thyroiditis is a painful, usually self-limited inflammation of the thyroid, often treated with the steroid prednisolone. A Zhejiang University hospital study looked back at 87 patients treated between 2013 and 2016 with prednisolone alone or prednisolone plus self-heal. This was a retrospective record review, not a randomised trial. Remission rates were similar; recurrence (4.4% with self-heal, 9.5% without) and temporary underactive thyroid (2.2% and 4.8%) did not differ significantly. Two mild skin rashes were recorded, both attributed to self-heal. The authors concluded that a lower-dose prednisolone regimen combined with self-heal was effective and safe.

Case reports

A 2021 report from Taiwan describes two people with Graves' disease, a 50-year-old woman and a 56-year-old man, who took the formula Jia Wei Xiao Yao San together with self-heal, fritillaria bulb and oyster shell. Their thyroid hormone levels returned to normal after 10 and 8 months and stayed controlled over a year of follow-up, with no side effects reported. Case reports describe individual patients; they cannot separate the effect of the herbs from the natural course of the disease or from other treatment.

Side Effects Recorded in the Trials

Side effects matter in thyroid treatment because the standard antithyroid drugs carry known risks, including rashes, liver injury and, rarely, a dangerous fall in white blood cells. The meta-analyses report adverse events as a total count rather than by type in their abstracts.

• Hyperthyroidism, 2025: adverse events were less frequent with self-heal added (risk ratio 0.34).
• Graves' disease, 2026: adverse events were less frequent with Prunella Oral Liquid added to antithyroid drugs (risk ratio 0.35, confidence interval 0.13 to 0.94).
• Thyroid nodules, 2021: no clear difference (risk ratio 0.67, confidence interval 0.36 to 1.22).
• Subacute thyroiditis, 2019: two mild skin rashes attributed to self-heal among the patients who took it.

The finding that adding a herb was followed by fewer side effects than the drug alone is unusual, since an extra treatment usually adds side effects rather than removing them. The abstracts do not explain the mechanism, and in open-label trials (where everyone knows who is taking what) side effects are among the outcomes most open to reporting bias. The reviews that report fewer adverse events also rate their overall evidence as low certainty.

Why the Reviewers Rate the Certainty as Low

Every review on this page reaches a favourable summary and a cautious verdict at the same time. Those are not contradictory: the numbers favour the add-on group, and the reviewers explain why the numbers may overstate the effect. The reasons they give are these.

1. Unclear allocation concealment. The 2026 Graves' disease review names this first. If the people enrolling patients can foresee which group the next patient will join, the groups can end up unequal.
2. No blinding. The same review names "performance bias due to lack of blinding". Without a placebo, patients and doctors knew who took the herb, which can shape symptom scores, side-effect reporting and composite "efficacy" ratings.
3. Low and very low GRADE ratings. The 2025 hyperthyroidism, 2026 Graves' disease and 2020 Hashimoto's reviews all graded the evidence low or very low.
4. "Uncertain" trial quality. The 2021 nodule review uses that word for the trials it included.
5. A single high-bias trial can sway the result. The 2020 Hashimoto's review says so of its own pooled estimate.
6. Calls for better trials. All four reviews end by asking for larger, more rigorous randomised trials; the 2026 review asks specifically for multicentre trials with standardised protocols.

Two further features are visible from the reviews' own methods. The trials were drawn largely from Chinese databases and come from one country, and the outcomes were mostly laboratory and ultrasound numbers rather than events patients notice, such as relapse after stopping medicine, the need for surgery or radioactive iodine, or quality of life. Where the 2025 review did measure relapse, it found no difference.

Laboratory and Animal Studies

The following results come from cells and animals. They suggest how the herb might act; they do not show that it does so in people.

• Goitre, in rats: in a 2026 study, 40 Wistar rats were fed an iodine-deficient diet for 12 weeks to enlarge the thyroid, then given levothyroxine or one of three doses of self-heal water extract for 16 weeks. In these rats, the medium and high doses reduced thyroid volume and inflammatory infiltration in the gland, with no detectable change in thyroid hormone levels or in the liver and kidney markers measured. The extract contained rosmarinic acid and salvianic acid A.
• Autoimmune thyroiditis, in rats: in Lewis rats given an injection of thyroglobulin to provoke a Hashimoto's-like disease, self-heal lowered thyroglobulin antibody levels, thyroid volume and inflammation scores, and expanded regulatory T cells, the immune cells that restrain autoimmunity. Rosmarinic acid was the main compound identified in the extract.
• Graves' disease, in mice: in female mice in which Graves'-like disease was induced with a virus carrying part of the TSH receptor, xia ku cao and its flavonoid luteolin lowered T4 and TRAb and reduced markers of oxidative stress.
• Human thyroid cells: in cultured thyrocytes (thyroid cells), self-heal reduced the inflammatory signals set off by viral-like genetic material, protected the cells from death, and turned down genes involved in iodide uptake and oxidation.
• Thyroid eye disease, in cells: self-heal polysaccharides at concentrations above 400 μg/mL slowed the growth of eye-socket connective-tissue cells taken from people with thyroid eye disease, but not cells from people without the disease.

Concentrations that act on cells in a dish are often far higher than anything reached in the body after swallowing a preparation, and how much of the herb's compounds reach the human thyroid has not been measured.

Iodine, Other Medicines and Safety Signs

The iodine question

Readers often ask whether self-heal acts on the thyroid by supplying iodine. None of the papers read for this page reports the iodine content of self-heal, so this page gives no figure. Two findings bear on the question indirectly: in the rat goitre study, the gland shrank without any detectable change in thyroid hormone levels, and in the thyroid-cell study, self-heal turned down iodide-uptake genes rather than supplying iodine. Iodine itself is covered on the Goiter page.

Other medicines

In the trials, self-heal was taken alongside antithyroid drugs, levothyroxine, radioactive iodine and prednisolone. None of the review abstracts reports a formal drug-interaction study, and no measurement of how the herb's compounds behave in the human body is reported in the papers read here. "No interaction reported" in these trials means no study looked for one, not that one has been ruled out. The Rosmarinic Acid, Safety and Unknowns page sets out the wider safety picture for the plant.

Signs hospitals treat as urgent

An overactive thyroid can tip into thyroid storm: high fever, a very fast or irregular heartbeat, agitation or confusion, vomiting or diarrhoea. Hospitals treat that combination as an emergency. On antithyroid drugs, fever with a sore throat is the sign hospitals check for a dangerous fall in white blood cells. A neck lump that is hard, fixed or growing is the presentation thyroid specialists investigate for cancer; a herb's traditional reputation for "dissolving nodules" does not tell a benign nodule from a malignant one.

What Is Not Known

1. Whether the effect survives blinding. No placebo-controlled, double-blind trial appears in the reviews' summaries.
2. Whether the effect appears outside China. The trials come from one country, mostly from Chinese-language journals.
3. Whether self-heal works on its own. The trials tested it as an add-on; they do not show what it does without standard medicine.
4. Whether patients feel or fare better. Lab numbers and ultrasound sizes changed; relapse, where measured in hyperthyroidism, did not.
5. Which dose and which preparation. The reviews pooled different preparations, and the abstracts do not give a single dose.
6. How the preparations are standardised. Oral liquids, granules and decoctions may differ in content.
7. The mechanism in people. Immune and anti-inflammatory mechanisms have been shown in rats, mice and cells only.
8. The iodine content of the herb, which no paper read here reports.
9. Interactions and long-term safety, which no trial abstract reports studying.

Decisions about thyroid medicines and any herb taken alongside them belong with a clinician who can follow the blood tests.

Key Research Papers

1. Wei M, Zhao Q, Yuan M, Fan Y, Li M. Efficacy and safety of Prunella vulgaris L. combined with antithyroid drugs for hyperthyroidism: a systematic review and meta-analysis. Front Pharmacol. 2025;16:1530152. PubMed PMID: 40083379
2. Chen Z, Yuan B, Zhang L, Zhang L, Yu F, Yu J, Yu H, Chen W. Prunella Oral Liquid in combination with Western medical treatment for Graves' disease: A systematic review and meta-analysis. J Ethnopharmacol. 2027;373:122255. PubMed PMID: 42526549
3. Han Q, Xu N, Chen B, Wu W, Sheng L. Safety and efficacy of Prunella vulgaris preparation in adjuvant treatment of thyroid nodules: A meta-analysis. Medicine (Baltimore). 2021;100(41):e27490. PubMed PMID: 34731129
4. Zhang YL, Hu RX, Zhao H, Yang W, Yu DD, Li HM, Liao X, Gu H. [Systematic review and trail sequential analysis of preparation of Xiakucao for Hashimoto's thyroiditis]. Zhongguo Zhong Yao Za Zhi. 2020;45(23):5777-5788. PubMed PMID: 33496119
5. Yang Z, Zhao N, Li J, Wu Z, Ma J. Effect of traditional Chinese medicine on Graves' disease: a network meta-analysis. Front Pharmacol. 2024;15:1411459. PubMed PMID: 39239642
6. Li SS, Chen SW, Gan D, Yu R, Gao TS. Efficacy and safety evaluation of traditional Chinese medicine formulations in treating graves' disease by improving immune function: A systematic review and meta-analysis. J Ethnopharmacol. 2026;366:121643. PubMed PMID: 41956233
7. Li F, Wu Y, Chen L, Hu L, Liu X. Initial treatment combined with Prunella vulgaris reduced prednisolone consumption for patients with subacute thyroiditis. Ann Transl Med. 2019;7(3):45. PubMed PMID: 30906749
8. Yang K, Guo KQ, Wu HY, Ye LX, Xia H. [Clinical effect of prunrllae oral solution in treating hyperthyrea]. Zhongguo Zhong Yao Za Zhi. 2007;32(16):1706-8. PubMed PMID: 18027674
9. Yang K, Guo KQ, Wu HY. [Clinical effect of Prunellae Oral Liquid on goiter with different thyroid function]. Zhongguo Zhong Xi Yi Jie He Za Zhi. 2007;27(1):37-9. PubMed PMID: 17302062
10. Lin CH, Lin CP, Huang ST. Successful intervention with Chinese herbal medicine for hyperthyroidism: Two case reports and a literature review. Explore (NY). 2021;17(4):344-350. PubMed PMID: 33109498
11. Han G, He X, Zhang C, Lai Y, Shan Z. Prunella vulgaris L. ameliorates goiter in Wistar rats is associated with changes in Rap1GAP/CREB3L1/PI3K-AKT signaling pathway. J Ethnopharmacol. 2026;374(Pt 2):122420. PubMed PMID: 42822614
12. Qiu H, Zhang J, Guo Q, Zhang Y, Zhong X. Prunella vulgaris L. attenuates experimental autoimmune thyroiditis by inducing indoleamine 2,3-dioxygenase 1 expression and regulatory T cell expansion. Biomed Pharmacother. 2020;128:110288. PubMed PMID: 32480225
13. Zhang Y, Qu X, Xu N, He H, Li Q, Wei X, Chen Y, Xu Y, Li X, Zhang R, Zhong R, Liu C, Xiang P, Zhu F. Mechanism of Prunella vulgaris L. and luteolin in restoring Tfh/Tfr balance and alleviating oxidative stress in Graves' disease. Phytomedicine. 2024;132:155818. PubMed PMID: 38879922
14. Chen F, Kawashima A, Luo Y, Kiriya M, Suzuki K. Innate Immune-Modulatory Activity of Prunella vulgaris in Thyrocytes Functions as a Potential Mechanism for Treating Hashimoto's Thyroiditis. Front Endocrinol (Lausanne). 2020;11:579648. PubMed PMID: 33304319
15. Li B, Guo J, Wang F, Cheng S, Zeng L. Effect of Prunella vulgaris polysaccharides on cultured orbit fibroblasts in vitro from patients with thyroid-associated ophthalmopathy. Exp Eye Res. 2020;201:108276. PubMed PMID: 32991886
16. Pan J, Wang H, Chen Y. Prunella vulgaris L. - A Review of its Ethnopharmacology, Phytochemistry, Quality Control and Pharmacological Effects. Front Pharmacol. 2022;13:903171. PubMed PMID: 35814234

PubMed Topic Searches

1. Prunella vulgaris and the thyroid
2. Prunella and Graves' disease
3. Prunella and Hashimoto's thyroiditis
4. Prunella and goitre

Connections

• Self-Heal (Prunella vulgaris) — the main topic page for the plant
• Self-Heal Benefits Deep Dive — the overview of every sub-article
• Rosmarinic Acid, Safety and Unknowns — the plant's wider safety record
• The Name and What It Promises — where the name xia ku cao comes from
• Graves' Disease — the most common cause of hyperthyroidism
• Hyperthyroidism — the overactive thyroid and its standard treatments
• Thyroid Nodules — how nodules are assessed
• Hashimoto's Thyroiditis — the autoimmune underactive thyroid

Featured Videos