We found no benefit of a chamomile extract, in the dose and formulation used, on our primary endpoints of subjective sleep efficiency and total sleep time in people with chronic primary insomnia in this preliminary study. We did, however, find a modest benefit of chamomile compared to placebo on other sleep diary measures including sleep latency (0.47 Cohen's d
), and night time awakenings (0.61 Cohen's d
), although these did not reach statistical significance. These measures of sleep quality are only slightly smaller than those observed for other drug treatments for chronic insomnia where pooled effect sizes for sleep diary measures (e.g., sleep latency, sleep quality, and wake after sleep onset) ranged from a mean of 0.38 to 0.79 [27
We observed that those in the chamomile group had a mean decreased sleep latency of a bit more than15 minutes, around 1/3 less night time awakening, and almost a 7% (4.2 point decrease) reduction in the FSS compared to placebo at day 28. These modest changes in sleep latency and night time awakenings are comparable with benzodiazepines, non-benzodiazepines and anti-depressants [27
]. Where weighted mean differences for decreasing sleep onset latency range from 7.0 to 19.6 minutes depending on the drug studied and how sleep latency was measured [27
]. This is despite our study patients, on average, experiencing milder insomnia than those recruited in other drug studies and thus, possibly attenuating our results due to having a less severely affected patient sample.
Despite the popularity of chamomile as a sleep aid [11
] there are no published studies examining chamomile for primary insomnia or for any type of chronic sleep disorder. Two other clinical studies have examined the sedative effect of chamomile as a secondary outcome in healthy volunteers and in women with menopausal symptoms [16
]. Somewhat similar to our results, both of these studies demonstrated that chamomile had positive effects on subjective sleep measures and in the case of menopausal women experiencing hot flashes improvements in fatigue as well. In contrast to our study, which found a mixed effect of chamomile on sleep measures, these studies both reported universally positive effects of chamomile on sleep parameters. However, sleep disturbances in menopausal women experiencing hot flashes may be caused by distinctly different mechanisms than those involved in primary insomnia. Also, it is unclear what improvements in sleep measures indicate in healthy volunteers without sleep disturbances.
Chamomile extract has also been investigated for treatment of stress and generalized anxiety disorders (GAD). Two small studies have evaluated the effect of chamomile essential oil on the autonomic nervous system. Masago and colleagues found that chamomile oil increased comfortable feelings and decreased alpha 1 (8-10 Hz) recordings of the EEG at the parietal and temporal brain regions [29
]. Chamomile oil versus placebo oil was also able to significantly shift negative mood images and frequency judgments in a positive direction after being asked to visualize positive or negative phrases [30
]. A RCT in people with mild to moderate GAD found that a chamomile extract was significantly better than a placebo extract at decreasing anxiety and showed a positive change in favor of chamomile for changes on the Psychological Well Being and Clinical Global Impression Severity Score [31
]. These studies indicate that chamomile may have mild anxiolytic activity, accounting for one mechanism by which it is an effective sleep aid. In our study, we purposely excluded individuals who had diagnoses of depression or anxiety in order to strictly evaluate the effect of chamomile on chronic primary insomnia. As such, we may have excluded the population most likely to experience sleep benefits from chamomile, i.e., those with insomnia that is related to an underlying anxiety disorder.
It is possible that our modest and mixed results were due to an adequate or incorrect; dose, dosing schedule or formulation of chamomile. There has been no dose finding studies with any chamomile product. Consequently, we based our dose on manufacturer recommendations and on doses of apigenin (chamomile's main flavonoid constituent), which produced sedative effects in animal models [12
]. Even if a higher dose of apigenin were found to improve sleep, it is not clear that it would be practical to deliver it using chamomile. In our current study our chamomile extract represented 15 gm of chamomile flowers/day. This amount could not feasibly be delivered as a tea and could only be achieved using a concentrated extract. Also, no studies have examined the pharmacokinetics or pharmacodynamics of chamomile, thus severely limiting the data of when and how often chamomile should be administered. Participants in our study took their tablets twice daily. Perhaps more frequent dosing or administration of chamomile through routes other than oral administration could yield better results. While pharmacokinetic studies of chamomile are lacking, other studies indicate that apigenin is orally available and has a reasonable half-life to justify twice daily dosing in humans when delivered in parsley (another rich source of apigenin) [35
]. Also, it may take longer than four weeks to see larger and more consistent effects of chamomile on sleep and fatigue measures. For instance, the study that found an effect in menopausal women was a 12-week RCT,[16
] although if positive effects take longer than one month many insomnia sufferers are likely to abandon the treatment. Moreover, the formulation could have been incorrect. For instance, drinking chamomile as a hot tea, or inhaling it as an essential could produce better results for improving sleep measures. Indeed, the other two studies that found a positive effect of chamomile on sleep measures used different formulations, a chamomile jelly and a tablet, combining chamomile with another herb Angelica sinensis
Chamomile appeared to be well tolerated. There was no difference between placebo and chamomile for all adverse events, for common AE categories including gastrointestinal complaints. No serious adverse events were reported and all non-serious adverse events were mild and transient in nature. Also, our participants were highly adherent taking more than 93% of their tablets on average, indicating that chamomile tablets are both safe and acceptable as a treatment.
Our study had several limitations. First, we were limited by the use of subjective, i.e., pen and pencil, sleep measures. Objective measures of sleep, such as actigraphy or polysomnography, would have been valuable to include and should be used in future studies. Second, we did not assess the characteristics of participants' sleep difficulties, i.e., initial, middle, late or early morning, mixed or non-restorative sleep. These differences in when participants are experiencing sleep issues may be important considering the age-range of individuals included in the study (20-65 years old) and thus, it could be possible that chamomile extract might be more effective for one type or another of sleep difficulties. We were unable to examine these differences in timing due our small sample size with only a few if any participants in any given time category. Also, this was a pilot study and thus we had a small sample size of only 34 participants. However, it was our intention to determine the effect size, direction of effect, and clinical relevance of chamomile on sleep and fatigue outcomes not necessarily statistical significance. Thus, our small number of participants was able to give a good indication of which sleep and fatigue parameters chamomile may benefit and guide future clinical trials in chronic insomnia sufferers. Also, we were able to generate data to determine how many participants would be needed to detect statistically significance differences. For instance, to detect a statistically significant difference in the sleep diary variable with the smallest effect size (0.06 for WASO), a total sample size of 2175 would be required, while only 60 participants (30 per group) would be necessary to detect a significant difference in sleep onset latency (effect size of 0.47).