Human beings spend around one-third of their lives sleeping^[1], which is more than 25 years of sleep, assuming an average human lifespan^[2]. Why is sleep so essential to justify spending so much time in this inattentive and motionless state? Unfortunately, there’s still no clear answer to that question. Despite decades of research, the underlying mechanisms and functions of sleep^[3] are still shrouded in mystery. Yet, some of the functions of sleep can be inferred from the effects of its deprivation. In all mammals, sleep deprivation leads to lack of attention, emotional instability^[4], heightened sensitivity to pain^[5], metabolic and cardiovascular disorders^[6], immune dysfunction^[7], and, in extreme cases, death^[8]. Sleep certainly seems to be essential for good brain function. The scientific literature has thus far counted about 10,000 genes expressed in the brain during sleep^[6]. It is therefore not surprising that sleep affects all aspects of our lives.

In large parts of the developed world, sleep deprivation seems to become a major problem^[9]. Poor sleep, in terms of either quantity or quality, is a risk factor for several illnesses, including obesity, cardiovascular^[10] and neurodegenerative diseases^[11], as well as metabolic^[12] and mental disorders^[13]. Moreover, lack of sleep is associated with higher all-cause mortality risk. Sleeping less than six hours per night is associated with a relative increase in mortality risk of 12%. This increased risk is carried by 9% of the population sleeping less than six hours per night in the U.S. On top of that, about one-third of Americans complain about sleep problems^[14].

But not all sleep problems are alike. Sleep problems can be broadly categorized into primary and secondary sleep disorders^[15]. Primary sleep disorders are caused by endogenous disturbances that directly affect the ability to sleep well. Primary sleep disorders are relatively common. For example, 5% of the French population suffers from specific sleep disorders^[16] such as insomnia, apnea, and narcolepsy. In contrast, secondary sleep disorders are caused by other illnesses such as depression, asthma, or arthritis. Secondary sleep disorders are thus far more common. For example, the National Institute of Mental Health (NIMH) estimated that 17.3 million U.S. adults, about 7% of the population, had at least one major depressive episode in 2017. These numbers sound depressing yet reflect a tragic reality: a lot of people are living with chronic sleep problems.

What can be done about this? There are two kinds of therapy designed to treat sleep problems. First, lifestyle interventions^[17] such as ensuring healthy nutrition, sufficient exercise, and good sleep hygiene can improve sleep quality. Cognitive therapy^[18], yoga^[19], and meditation^[20] are also recommended. However, lifestyle interventions may not be suitable or effective for people suffering from serious diseases. In such cases, the second type of intervention, pharmacotherapy^[21], or “sleeping pills,” seems more promising. The American Academy of Sleep Medicine^[22] offers and recommends a plethora of prescription drugs. However, these medications have substantial side effects^[23], such as daytime sleepiness, cognitive impairment, dependency, and withdrawal. Due to these common adverse effects, researchers are looking for new promising treatment options. Melatonin is one of them.

Melatonin^[24] is a hormone and neurotransmitter that plays a crucial role in regulating sleep. Melatonin is secreted by the^[25] brain’s pineal gland (shown in Figure 1) in the evening hours^[26], approximately two hours before^[27] bedtime. However, in our modern times, nocturnal melatonin secretion can be disturbed by excessive exposure to artificial lights^[28] such as smartphones^[29], televisions^[30], and indoor lighting. Furthermore, certain illnesses, such as depression^[31], can negatively impact melatonin secretion. Thus, supplementation with exogenous melatonin could improve sleep quality.

Figure 1: Melatonin 101

The biological theory of melatonin’s usefulness has been backed by rapidly accumulating clinical evidence. The three most recent meta-analyses from 2015^[32], 2016^[33], and 2018^[34] examined the efficacy of melatonin on primary and secondary sleep disorders. All three meta-analyses found that melatonin improved sleep quality significantly as assessed by the global PSQI score, total sleep time, and sleep latency onset, which is how long it takes to fall asleep. The current body of literature thus supports the use of melatonin to treat sleep disorders.

At first glance, the clinical evidence seems convincing. There are, however, three remaining caveats with melatonin. First, the observed effects were significant but small, and the results were often measured with suboptimal measures that assess only one or two dimensions of sleep quality. Second, the number of available RCTs is still low, and the existing studies often have short follow-ups and insufficient sample sizes. Third, the existing meta-analyses did not differentiate the different diseases that cause secondary sleep disorders. As a result, the American Academy of Sleep Medicine^[22] guidelines and the European insomnia guidelines do not recommend melatonin to treat sleep disorders based on the current evidence. In other words, the evidence exists, but is not yet convincing enough.

The present study^[35] was designed to fill this gap. The researchers meta-analyzed 23 RCTs that examined melatonin’s effect on sleep quality in 1,965 participants with different diseases. In contrast to previous meta-analyses, the present study focuses solely on RCTs that used the Pittsburgh Sleep Quality Index (PSQI), the most widely used and reliable^[36] subjective measure of sleep quality to date. The researchers also conducted subgroup analyses to assess whether melatonin is more effective in the context of some diseases than in others.

Sleep is essential for life, and insufficient sleep is a risk factor for several illnesses. As the hormone and neurotransmitter melatonin regulates sleep, supplementation may improve sleep quality. Due to insufficient evidence, however, professional societies are not recommending melatonin to treat sleep disorders yet. The present study was designed to meta-analyze the most recent high-quality evidence from RCTs to fill this gap.

The present study is a systematic review and meta-analysis of 23 RCTs published from 2004 to June 2020 that investigated melatonin’s effects on the sleep quality of adults with various diseases. Notably, the researchers only included RCTs that used the Pittsburgh Sleep Quality Index (PSQI)^[37], the most widely used subjective measure of sleep quality (see the sidebar below). Of the 23 included RCTs, 3 studies were cross-over trials, while the remaining 20 RCTs used a parallel study design. The sample size ranged from 16 to 711 participants per study, with follow-ups of 2–24 weeks. The included RCTs used either melatonin or prolonged-release melatonin as an intervention with doses of 2–10 mg. RCTs that lacked sufficient data for the outcomes of interest and studies carried out with less than two weeks of follow-up were excluded. Overall, the pooled sample size was 1,965 participants from 13 countries.

The researchers used a random-effects model^[40] for the meta-analysis and reported their findings based on the Preferred Reporting Items of Systematic Reviews and Meta-Analysis (PRISMA)^[41] statement guideline. The researchers also conducted subgroup analyses to delineate the effects of melatonin on different diseases and interventions. Among others, they included participants with respiratory and neurodegenerative diseases, as well as metabolic, sleep, and mental disorders. Notably, the researchers did not preregister the meta-analysis. To evaluate the quality of the included RCTs, the researchers used the Cochrane scoring system^[42] to assess the risk of bias (high risk, low risk, or unclear risk). The risk of publication bias was judged by visual inspection of funnel plots.

The present study is a meta-analysis of 23 RCTs investigating the effects of melatonin on sleep quality in 16,711 adults with various diseases. Unlike previous meta-analyses, the authors restricted included only RCTs that assessed sleep quality using the Pittsburgh Sleep Quality Index (PSQI) score—the most widely used, validated subjective measure of sleep quality. Subgroup analysis was used to examine the effect of melatonin on different diseases and kinds of interventions.

Overall, melatonin did significantly improve sleep quality in adults with various diseases. Sleep quality was assessed using the PSQI score, ranging from 0 to 21 with a score of 0–4 indicating good sleep quality and a score of more than 4 suggesting poor sleep quality. The present meta-analysis found a change in global PSQI score of -1.24. It’s difficult to put this number into context, though, since baseline scores were not reported. However, the author of the PQSI has stated^[43] that a change of 3 or more indicates a clinically important effect.

However, the researchers found considerable heterogeneity in the data set, meaning the RCTs used for the meta-analysis were very different. This heterogeneity notably could not be explained by the overall risk of bias, which was low, and the study quality, which was overall good. Hence, the researchers aimed to identify heterogeneity sources by performing subgroup analyses on three aspects: health status, kind of intervention, and other factors.

The subgroup analyses, shown in Figure 2, revealed that melatonin significantly improved the sleep quality (PSQI score) of participants with respiratory diseases (-2.2), sleep disorders (-0.67), and metabolic disorders (-2.74). In contrast, participants with mental, neurodegenerative, and other conditions did not experience an improved sleep quality when taking melatonin. The researchers found that both melatonin (-1.52) vs. prolonged-release melatonin (-0.71) significantly improved sleep quality (PSQI score). Even though melatonin showed twice the effect of prolonged-release melatonin, the difference was not significant. In other words, the kind of intervention did not make any statistically significant difference in terms of helping people sleep better. Baseline BMI, doses of melatonin (2–10 mg), study duration (2–24 weeks), age, and gender were examined by subgroup analysis. The researchers found no significant differences for these factors.

Figure 2: Results with 95% confidence intervals

Melatonin significantly improved sleep quality in adults with various diseases by a PSQI score of -1.24. It’s unclear, however, if this small change in PSQI score, which ranges from 0 (good sleep) to 21 (poor sleep), may be clinically relevant. This effect was independent of doses and melatonin formulations, but the data set was considerably heterogeneous. The researchers found a reason for this heterogeneity in the large number of different diseases included in the 23 RCTs. As shown by subgroup analysis, melatonin significantly improved sleep quality for mental respiratory diseases as well as sleep and metabolic disorders. In contrast, the subgroup analysis found no effects of melatonin on participants with cognitive, neurodegenerative, and other conditions.

The present meta-analysis provides valuable insights into the effects of melatonin on secondary sleeping disorders. Most importantly, the study only included RCTs using the PSQI score (the most reliable subjective measure of assessing sleep quality) and analyzed the effects of melatonin on sleep in people in different disease states separately via subgroup analyses. However, the present study’s findings do not exist in a vacuum, so they require context for interpretation. In particular, three aspects need to be addressed: the validity of the PSQI score, comparing the present study to previous meta-analyses, and discussing the varying effectiveness of melatonin on different diseases.

The present study’s researchers only included RCTs that assessed sleep quality using the Pittsburgh Sleep Quality Index (PSQI)^[37]. It’s reasonable to ask: Why exclude all other studies? Practically speaking, there are good reasons to rely on the PSQI as a sole measure of sleep quality.

The PSQI is a short self-questionnaire that is practical and easy to use. In contrast, objective measures such as polysomnography would be more accurate and precise but also more time-consuming and cost-expensive, difficult to interpret, and challenging to use in large-scale clinical trials. Furthermore, the validity and reliability of the PSQI have been independently validated by numerous studies^[38] (see sidebar).

As the PSQI measures seven dimensions of sleep, it would be interesting to know which dimensions are influenced most by melatonin. Unfortunately, the present study’s researchers do not report on the individual items, only the global PSQI score. This information would, however, be helpful when interpreting the results. Previous meta-analyses^[34][33][32] repeatedly showed that melatonin primarily affects sleep onset latency, i.e., the time it takes to fall asleep. However, this has only been demonstrated for pooled data sets and not in (sub-)analyses restricted to participants with specific diseases. Future studies should address this open question.

In contrast to previous meta-analyses, the present study differed in two crucial ways (see Figure 3). First, the present study is more comprehensive in terms of included studies (23 RCTs) and sample size (1,965 participants). This sample size is almost four times larger than the meta-analysis on secondary sleep disorders from 2015^[32] and twice the size compared to the largest meta-analysis on primary sleep disorders^[33] from 2016. The comprehensive and diverse data set also enabled the researchers to perform subgroup analyses, discussed below.

The present meta-analysis included only RCTs that examined secondary sleep disorders using the PSQI method. Previously, only one other meta-analysis^[32] used the PSQI score to assess melatonin’s effects on Alzheimer’s and Parkinson’s disease-related sleep disorders. This study found a -4.2 lower PSQI score for Alzheimer’s disease, but no change in global PSQI score for Parkinson’s disease. In contrast, the present study found a non-significant effect of melatonin (PSQI -1.2) for neurodegenerative diseases. What explains this outcome discrepancy?

Heterogeneity is most likely the root cause for the discrepancy. The present study pooled four studies into one subgroup (one RCT on Alzheimer's and three RCTs on Parkinson’s). Only one study^[44] showed a significant improvement of melatonin on sleep, while the other three did not. As a result, statistical power was low, and the subgroup as a whole did not show any significant effect. Overall, it seems that melatonin is more effective in treating sleep disorders in Alzheimer’s than Parkinson’s disease patients.

However, based on the mechanism of action proposed for melatonin, it’s reasonable to hypothesize that its beneficial effects on sleep quality would be independent of the disease. However, the present study contradicts this hypothesis. Melatonin significantly improved the sleep quality only in participants suffering from respiratory diseases, metabolic disorders, and sleep disorders. In contrast, participants with mental, neurodegenerative, and other conditions did not experience an improved sleep quality when taking melatonin. The authors, unfortunately, do not comment on this issue. Here are two possible explanations.

First, melatonin may improve only specific aspects of sleep that are only relevant for some diseases. Indeed, it has been repeatedly shown that melatonin primarily affects sleep onset latency and with that also total sleep time to some extent. Other aspects of sleep may remain unaffected, though. One reason for this may be that melatonin is rapidly metabolized and removed from systemic circulation after secretion or, in the case of supplementation, exogenous intake. In contrast, certain diseases may disrupt specific aspects of sleep that melatonin does not influence. As a result, melatonin’s beneficial effects on one aspect of sleep would be disguised by the probably much higher negative effects the disease has on another aspect. In other words, sleep disorders are complex and go way beyond insufficient or impaired release of melatonin. More research is needed to unravel the underlying mechanisms of melatonin and sleep disorders.

The second explanation is more of a statistical nature: Melatonin may improve sleep for all diseases, but the available data quality could be insufficient to detect these effects. Researchers need data in sufficient quantity and adequate quality to observe these effects with high accuracy and precision. Unfortunately, this was not the case for the present study, which could explain why melatonin improved sleep quality only for three of six studied diseases. Even though the current meta-analysis counts almost 2,000 participants, some subgroups consisted of only a few studies and few participants. Thus, these subgroups lack statistical power and the findings could therefore be false negatives. They also may not represent the entire population affected by these sleep disorders. One observation in support of this criticism is that the three diseases for which melatonin did not show a significant effect showed the highest heterogeneity. More research on these diseases, their effects on sleep and potential modulatory effects of melatonin, is required.

The present study is one of the most extensive meta-analyses on melatonin and sleep quality in participants with various diseases to date. Three aspects are particularly relevant here: First, the present study’s findings are in broad agreement with previous research. Second, the inclusion criterion to select only RCTs that used the PSQI to assess sleep quality seems justified and further added to the quality of the present meta-analysis. Third, subgroup analyses, though methodologically questionable due to the low number of studies and participants, provided a more nuanced picture of how melatonin influences sleep quality in different diseases. The disease subgroups that did not show any significant effect of melatonin on sleep quality were all highly heterogeneous. More research is required to examine whether melatonin could improve sleep quality in participants suffering from specific diseases. However, it can be said that melatonin slightly improves sleep quality in the context of respiratory, mental, and sleep disorders. From a mechanistic point of view, it should be beneficial for other conditions, too, but a verdict on their usefulness beyond the three previously mentioned disease states will have to wait for more evidence until more evidence becomes available.

Q. Can melatonin help with jet lag or shift work?

Yes. The current evidence suggests that melatonin can ease sleep problems associated with jet lag. In situations where external stimuli (sunlight and darkness cycles) and internal stimuli (the internal clock) are not in sync, supplemental melatonin is thought to help re-establish balance. A Cochrane meta-analysis of 10 studies^[45] that examined jet-lag from flights that transversed five or more time zones found that melatonin was significantly more effective than placebo, when taken at the destination’s bedtime, for normalizing the circadian rhythm and reducing the symptoms of jet lag.

According to the reviewed studies, there is no significant difference between doses of 500 mcg and 5 mg on the effects of melatonin in terms of reducing jet lag. The researchers noted some better sleep with the 5 mg dose. Note that some people still experienced jet lag, as the meta-analysis noted that in the two studies that reported individual statistics, about 18% of participants still experienced jet lag after melatonin, with placebo at 67%.

Overall, melatonin seems to be helpful to re-establish a healthy circadian rhythm after experiencing jet lag. Interestingly, melatonin could also be used to treat social jet lag^[46]. Future studies may shed more light on this.

Q. Is there any evidence regarding melatonin toxicity, tolerance, and addiction?

Overall, melatonin supplementation is relatively safe. First, melatonin toxicity is very low. Studies testing up to 240 mg^[47] and 500 mg^[48] orally taken do not report any toxic effects. Second, tolerance does not seem to be an issue. A study^[49] examining the continued administration of melatonin over 6–12 months did not find any evidence of tolerance. Third, there is currently no evidence to support addiction or withdrawal symptoms. However, doses higher than 2 mg have not been sufficiently studied over the long term. Lastly, if you’re worried about the potentially testosterone-lowering effect of melatonin, rest assured, there is currently no conclusive evidence to support this claim.

Sleep is vital for maintaining well-being. However, many people experience chronic sleep problems, either from primary or secondary sleep disorders. Lifestyle interventions and prescription drugs are commonly used to treat these sleep disorders. However, these strategies, especially pharmaceuticals, tend to have severe side effects. Thus, alternative treatment options with fewer adverse effects are being investigated. Melatonin is a hormone and neurotransmitter that has caught the attention of sleep researchers due to its high safety and tolerability. However, practical guidelines from leading associations still do not recommend melatonin to treat sleep disorders due to insufficient evidence.

The present meta-analysis was designed to address this issue and examined the effects of melatonin on sleep quality in participants with different diseases. The researchers meta-analyzed 23 RCTs that assessed sleep quality using the Pittsburgh Sleep Quality Index (PSQI). Overall, the data showed that melatonin improved sleep quality significantly, though the researchers also found considerable heterogeneity. Health status was found as a cause of this heterogeneity. Subsequent subgroup analysis revealed that melatonin significantly improved sleep quality in participants with respiratory diseases, metabolic disorders, and sleep disorders, but not with neurodegenerative, mental, or other conditions. Notably, the latter three diseases’ subgroups were highly heterogeneous, possibly explaining the lack of statistical significance. More research is required to examine if melatonin improves sleep quality in the context of these diseases.