Herbal Teas Compared: What the Research Shows and How to Brew Each One

A meta-analysis of 12 randomized trials found a significant improvement in sleep quality with chamomile (standardized mean difference -0.73, 95% CI -1.23 to -0.23) and a significant improvement in generalized anxiety disorder scores on the HAM-A scale after 2 and 4 weeks. It found no difference for state anxiety (three trials), and the one trial on insomnia found no significant change. The authors called for larger trials.

A 15-day open-label pilot gave 20 volunteers with mild-to-moderate anxiety and sleep disturbances a standardized lemon balm extract. Clinician ratings showed reductions in anxiety (18%) and insomnia (42%). With no placebo group, the study cannot separate the extract from expectation. A 2024 review concluded lemon balm holds promise but that robust randomized trials are required.

A systematic review identified 24 human studies of tulsi, and all reported favorable clinical outcomes (metabolic, cardiovascular, immune and neurocognitive) with no significant adverse events. The review searched books, theses and conference proceedings as well as journals, and it called for more research on dose and form. One controlled trial it describes gave 150 adults with stress a whole-plant capsule (400 mg, three times a day) for 6 weeks and reported fewer stress-related symptoms than placebo, including fatigue and sleep problems.

A meta-analysis of 10 randomized trials (1,030 patients) found peppermint oil better than placebo for global irritable bowel syndrome symptoms (relative risk of not improving 0.65; number needed to treat 4) and for abdominal pain (relative risk 0.76; number needed to treat 7). Adverse events were more common with peppermint oil (relative risk 1.57). These were trials of peppermint oil, and the analysis does not address peppermint tea.

A review of meta-analyses reported that ginger supplementation was associated with significant reductions in C-reactive protein and tumor necrosis factor-α. In pregnant women, ginger significantly eased nausea but had no significant effect on how often vomiting occurred. Doses were typically 1 to 3 g a day for the inflammation outcomes and 500 to 1,500 mg a day in divided doses for pregnancy nausea. The authors noted high heterogeneity and risk of bias in some included studies.

A systematic review of 23 randomized trials in people receiving chemotherapy found no significant difference between ginger and control in the incidence of acute nausea, delayed nausea, acute vomiting or delayed vomiting overall. In a subgroup taking no more than 1 g a day for more than four days, acute vomiting was lower (odds ratio 0.30; 95% CI 0.12 to 0.79). The authors concluded ginger may reduce acute vomiting but that the evidence could not confirm an effect on nausea or delayed vomiting.

In a randomized, double-blind, placebo-controlled trial of 312 air travelers, a standardized elderberry extract did not significantly change the number of cold episodes (12 versus 17, p = 0.4). The placebo group had significantly more cold-episode days (117 versus 57, p = 0.02) and a higher average symptom score (583 versus 247, p = 0.05). The authors described this as a significant reduction in cold duration and severity and asked for more research.

A meta-analysis of randomized trials of elderberry supplementation (180 participants in total) found a large pooled reduction in upper respiratory symptoms. The trials were small, and they tested elderberry supplements rather than tea.

The Cochrane review pooled 24 double-blind trials (4,631 participants) of echinacea and did not pool the main results because the products differed so much. None of the 12 prevention comparisons found a statistically significant difference, although a post hoc pooling suggested a 10 to 20% lower relative risk of a cold. Of seven treatment trials reporting cold duration, only one showed a significant effect, and the authors concluded echinacea products had not been shown to benefit treatment of colds. An earlier meta-analysis of 14 studies (Shah 2007) reached a more positive result, with 58% lower odds of a cold and colds 1.4 days shorter. The two reviews disagree, and the products tested varied widely.

In a randomized, double-blind, placebo-controlled trial, 74 people with allergic rhinitis took a 150 mg nettle root extract tablet or a placebo for a month, and 40 completed it. Symptom scores improved significantly in both groups, so placebo produced a similar improvement, while nasal eosinophil counts fell with nettle. The authors called for larger, longer trials. This trial used a root extract and did not test nettle leaf tea.

A meta-analysis of 17 randomized trials found hibiscus lowered systolic blood pressure by an average of 7.10 mmHg compared with placebo (95% CI -13.00 to -1.20), with very large variation between trials (I² = 95%). The reduction was greatest in people with elevated blood pressure at the start. Compared with medication, hibiscus produced reductions that were not significantly different from the drugs. Low-density lipoprotein fell by 6.76 mg/dL compared with other teas and placebo (95% CI -13.45 to -0.07; P = 0.05). The authors said dose and treatment duration still need to be established.

A scoping review of rooibos tea in people covered 18 publications with 488 participants in total, with half of the studies conducted in South Africa and doses from 200 to 1,200 ml. It reported improved lipid profiles, better antioxidant status and lower blood glucose in healthy and at-risk people, while noting that the studies were few and small, the types and doses of tea varied, and results were inconsistent. Most of what is known about rooibos comes from animal and cell studies.

In healthy volunteers given 2 g of curcumin, serum levels were either undetectable or very low. Adding 20 mg of piperine, the compound in black pepper, raised the bioavailability of curcumin by 2,000%. This is why turmeric drinks such as golden milk are traditionally made with black pepper and a fat source.

A systematic review of 29 randomized trials (3,846 participants) of silymarin supplements, at doses from 140 to 420 mg, found 65.5% of studies reported lower liver enzyme levels, 20.7% found no significant change and 13.8% reported higher enzyme levels. The authors described the results as context-dependent.

The National Center for Complementary and Integrative Health states that there is not enough high-quality evidence to clearly support the use of licorice for any health condition. Some studies suggest a licorice mouth rinse or gargle might reduce ulcer size and pain in recurrent canker sores, and a gargle before surgery might lessen sore throat after intubation. These were rinses and gargles, not tea. In a 2024 randomized, non-blinded crossover trial of 28 healthy adults, two weeks of daily licorice containing 100 mg of glycyrrhizic acid, an amount the World Health Organization had suggested was unlikely to cause adverse effects, raised home systolic blood pressure by 3.1 mmHg compared with control and suppressed renin and aldosterone. The authors suggested the safe limit might need to be reconsidered.