Melatonin (N-acetyl-5-methoxytryptamine) is a neurohormone produced by pinealocytes in the pineal gland during the dark hours of the day-night cycle. Serum levels of melatonin are very low during most of the day, and it has been labeled the “hormone of darkness.” Melatonin is involved in the induction of sleep, may play a role in the internal synchronization of the mammalian circadian system, and may serve as a marker of the “biologic clock” (Haimov & Lavie, 1995; Garfinkel et al, 1995; Dollins et al, 1994; Tzischinsky & Lavie, 1994; Jan et al, 1994; Cavallo, 1993; Short, 1993).
Actions & Pharmacology
Melatonin is thought to have antioxidant, immunomodulator, and hypnotic effects.
In general, the pineal gland (projecting from diencephalon into third ventricle) is a neuroendocrine transducer, related to its secretion of melatonin. The hormone serves as a messenger to the neuroendocrine system regarding environmental conditions (especially the photoperiod). Putative functions of endogenous melatonin in this regard include regulation of sleep cycles, hormonal rhythms, and body temperature (Deacon et al, 1994; Dollins et al, 1993; Cavallo, 1993). Melatonin may also have a role in influencing the maturation and function of the hypothalamic-pituitary-gonadal axis and in determining the onset of puberty (Cavallo, 1993).
Production of melatonin is regulated by postsynaptic receptors originating in the superior cervical ganglion, which innervate the pineal gland. The suprachiasmatic nucleus of the hypothalamus (entrained by the light-dark cycle and considered the anatomic site for the biologic clock) receives stimuli from the retina (retinohypothalamic tract), and during dark hours the suprachiasmatic nuclei forward a stimulus to the superior cervical ganglion and pineal gland, resulting in melatonin secretion (Haimov & Lavie, 1995; Cavallo, 1993). This stimulatory activity is suppressed by light, especially bright light (Thalen et al, 1995; Cavallo, 1993; Strassman et al, 1987). Melatonin synthesis in the pinealocyte is dependent upon noradrenergic stimulation (Cavallo, 1993). The normal endogenous production rate is 28 to 30 mcg/d (Short, 1993; Lane & Moss, 1985). Production of the hormone is reduced in cirrhotic patients (12 mcg/day) (Lane & Moss, 1985) and in the elderly (Garfinkel et al, 1995).
Melatonin possesses free-radical scavenging properties, protecting cells against many oxidative agents. The mechanism of free-radical scavenging action is not well elucidated. Melatonin reacts with nitric oxide in the presence of doublet oxygen, producing N-nitrosomelatonin. Since nitric oxide can cause some cellular destruction, melatonin may protect the cell against such oxidative damage (Turjanski et al, 2000). Melatonin also inhibits the production of nitric oxide by inhibiting nitric oxide synthase. Nitric oxide's damaging effects are mediated through its reaction with superoxide to form peroxynitrite. Peroxynitrite stimulates lipid peroxidation, inactivates various enzymes, and depletes glutathione (Cuzzocrea et al, 2000).
In vitro and animal studies have reported that melatonin is capable of inducing direct cytostatic actions on some human cancer cell lines, stimulating host immune responses, and inhibiting release of somatomedin-C (Lissoni et al, 1995; Lissoni et al, 1991). Melatonin has been used alone and in combination with interleukin-2 as an immunotherapeutic regimen in the treatment of cancer (Lissoni et al, 1995; Lissoni et al, 1994).
Prolonged administration of oral melatonin has reportedly induced phase-setting effects on circadian rhythms, such as the sleep-wake cycle and rest-activity. The hormone has been reported to produce re-entrainment of circadian rhythms after time-zone shifts, and entrainment of previously free-running rhythms in the blind (Dollins et al, 1993; Cavallo, 1993; Dahlitz et al, 1991; Arendt et al, 1988; Arendt et al, 1986).
Findings from several studies have established a link between low serum levels of melatonin and high cancer risk, through the disruption of circadian rhythms. One update to the literature examined several studies of melatonin in the treatment of advanced cancer. In one randomized, controlled trial mentioned involving 30 patients with colorectal cancer, subjects received treatment with irinotecan alone (n=16) or with concomitant oral melatonin (20 mg/day, administered in the evening). A 4% disease control rate was observed in those receiving melatonin in addition to chemotherapy (12/14 vs 7/16, p<0.05). Overall, most studies of exogenous melatonin, given alone or in combination, showed increased survival rates and at least a partial response to therapy (Mahmoud et al 2005).
A systematic review examined 10 randomized trials published between 1992-2003 involving 643 subjects with solid tumors. Researchers investigated the effects of melatonin, given at bedtime, on treatment of cancer or supportive care and 1-year survival rates. Dosages ranged from 10-40 mg/day across all studies. In addition to being well tolerated, melatonin diminished the risk of mortality at 1 year, without severe adverse events (Mills et al 2005).
Tumor regression rate and 5-year survival results were significantly increased in metastatic non-small cell lung cancer subjects who were concomitantly treated with chemotherapy and melatonin. One hundred subjects were randomized to receive chemotherapy only or chemotherapy and melatonin. Chemotherapy subjects received cisplatin 20 mg/m2 per day intravenously and etoposide 100 mg/m2 per day intravenously for 3 consecutive days. Four chemotherapy cycles were planned at 21-day intervals. Melatonin 20 mg orally was given 7 days every evening before chemotherapy and was continued after chemotherapy. Toxicity and clinical response were evaluated according to criteria of the World Health Organization. A complete response (CR) was achieved in 2 of 49 (4%) subjects treated with chemotherapy and melatonin. No patient achieved a CR with only chemotherapy. Fifteen of 49 (31%) subjects treated with chemotherapy and melatonin achieved a partial response (PR) compared to 9 of 51 (18%) chemotherapy-only subjects. Disease progression occurred in 20 of 51 chemotherapy-only subjects while 6 of 49 chemotherapy-melatonin subjects had progression (p<0.01). At 5 years, chemotherapy plus melatonin compared to chemotherapy alone improved survival (3 vs 0, p<0.001). Chemotherapy treatment was also better tolerated in subjects treated concomitantly with melatonin (Lissoni et al, 2003).
Sixteen patients with glioblastoma who were treated with melatonin 20 mg/d and radiotherapy (RT) experienced prolonged overall survival time compared to 14 patients given RT alone. Both groups were given steroids and anticonvulsants. The melatonin-treated group had a higher rate of survival at 1 year than did the RT-only group (p<0.02). Patients with RT alone experienced a significantly higher number of infections compared to melatonin plus RT group (p<0.025) (Lissoni et al, 1996a).
Tamoxifen and melatonin administered to 25 patients with metastatic solid tumors showed beneficial effects in terms of controlling cancer-cell proliferation or improving performance status. Patients included those diagnosed with melanoma, uterine cervix carcinoma, pancreatic cancer, heptocarcinoma, ovarian cancer, small cell cancer, or unknown primary tumor. Tamoxifen 20 mg and melatonin 20 mg were given daily. In 3 patients (12%) a partial response was seen. Stable disease was observed in 13 patients (52%) with the remaining 9 patients demonstrating progressive disease. No toxicity was seen (Lissoni et al, 1996b).
Other studies have reported significantly prolonged survival and greater improvement in performance status with oral melatonin plus supportive care compared to supportive care alone in patients with non-small cell lung cancer (n=63) (Lissoni et al, 1992a) and brain metastases of solid tumors (n=50) (Lissoni et al, 1994a). In the non-small cell lung cancer patients, a dose of 10 mg daily for 21 of 28 days was administered. No complete or partial responses were observed, although stable disease was achieved in significantly more patients treated with melatonin (32% versus 9%). A dose of 20 mg daily until progression was given to patients with brain metastases; the free-from-progression period was greater and the frequency of steroid-induced metabolic and infective complications was significantly lower with melatonin therapy relative to supportive care alone in this study. In both studies, patients had failed or progressed on prior chemotherapy, although details of previous therapy or criteria for failure were not provided. Methods of randomization and pretreatment clinical status of patients (such as underlying conditions) in each group, which could affect outcome, were also not specified, and the numbers of patients may have been too small for adequate statistical analysis. The same group of investigators conducted all studies.
The combination of subcutaneous recombinant interleukin-2 (aldesleukin) given 5 or 6 days/week for 4 weeks, plus oral melatonin 10 to 50 mg daily produced complete or partial tumor responses in 23% of pretreated patients with various digestive tract tumors (ie, colorectal, gastric, hepatic, or pancreatic carcinoma) (Lissoni et al, 1993a), 21% with solely metastatic gastric carcinoma and low performance status (Lissoni et al, 1993), and 20% of patients with non-small cell lung cancer (first-line therapy) (Lissoni et al, 1992) in small uncontrolled studies. A partial response rate was also observed with this combination in 3 of 12 previously treated or untreated patients (25%) with locally unresectable or metastatic endocrine tumors; responses occurred in carcinoid tumor, neuroendocrine lung tumor, and pancreatic islet cell tumor (Lissoni et al, 1995).
Evening administration of melatonin throughout chemotherapy treatment and every day thereafter resulted in significant reductions in manifestations of chemotherapy-induced toxicity. Eighty patients were given melatonin 20 mg/d concomitantly with chemotherapy or chemotherapy only. Supportive care was the same in both groups of patients. Thrombocytopenia was significantly less (p<0.006) in the group receiving melatonin. Leukopenia and anemia were also less frequent. Asthenia and malaise were significantly less frequent (p<0.006) in the melatonin group. Stomatitis, neuropathy, and cardiac complications also occurred less frequently in the group receiving concomitant melatonin than in the group receiving chemotherapy alone. There was no difference between groups in the frequency of alopecia, nausea and vomiting, or diarrhea. The antitumor activity of cytotoxic drugs was not negatively influenced by concomitant administration of melatonin (Lissoni et al, 1997b).
Melatonin reduced the number of daily attacks of cluster headache and the consumption of analgesics in acute sufferers. In a double-blind study of 20 sufferers of cluster headaches, subjects were given either melatonin 10 mg in a single evening dose or placebo for 2 weeks, during a cluster period. Fifty percent of melatonin-treated patients responded. Improvement began 3 days after the onset of treatment, and responders were free of headaches after 5 days. There was no improvement in the placebo group. Discontinuation of treatment by responders was followed by gradual recurrence of cluster attacks, obliging re-institution of treatment (Leone & Bussone, 1998; Leone et al, 1996).
Melatonin eliminated the occurrence of headaches of various kinds (migraine, cluster, tension headache) in 5 patients with delayed sleep phase syndrome. Each patient received either melatonin 5 mg or a placebo nightly for 14 days, followed by the other treatment for 14 days. Thereafter, each patient took melatonin nightly for at least 3 months. Melatonin was given 5 hours before the time when endogenous melatonin reached 10 picograms/mL. The endogenous nocturnal melatonin level of all subjects was normal and did not differ from those of patients who had delayed sleep phase syndrome without headache, suggesting that the headaches were not caused by melatonin deficiency. Rather, relief from headaches may have been a result of increased sleep and synchronicity of the biological clock to lifestyle (Nagtegaal et al, 1998).
Jet Lag/Sleep Restriction
A meta-analysis examining the effect of exogenous melatonin on sleep onset latency in those with sleep disorders accompanying sleep restriction found no evidence of efficacy with supplementation. Nine randomized trials involving 427 participants were studied. However, two systematic reviews cited in the analysis, including a large Cochrane review, concluded that melatonin was effective in alleviating jet lag. This disparity may have been the result of an aspect overlooked in the meta-analysis—namely, measures of daytime fatigue—which was examined in the studies reporting a benefit with melatonin (Buscemi et al 2006). Another study used exogenous melatonin administration combined with exposure to light and physical exercise to hasten resynchronization after transmeridian flights (comprising 12-13 time zones). Improvements were noted after about 2 days of treatment, compared with an average resolution occurring over 8 to 10 days following such a flight (Cardinali et al 2006).
In some earlier jet lag studies, melatonin was not statistically superior to placebo (Claustrat et al, 1992; Petrie et al, 1989). In particular, mood, sleep quality, and morning sleepiness were not altered significantly in one study (Claustrat et al, 1992). Effective doses of melatonin have been either 5 mg daily (at various times) for 3 days prior to the flight, then the same dose for 4 additional days, or an 8 mg dose on the day of the flight and for 3 further days. The former regimen appears more effective.
A controlled-release formulation of melatonin (2 mg) administered nightly reduced nocturnal systolic and diastolic blood pressure in patients with confirmed nocturnal hypertension. The double-blind study randomized patients to treatment with melatonin 2 hours before bedtime (n=19) or placebo (n=19) for 4 weeks. Melatonin treatment significantly reduced nocturnal systolic BP from 136±9 to 130±10 mm Hg (p=.011) and diastolic BP from 72±11 to 69±9 mm Hg (p=.002). No effect on nocturnal blood pressure was observed with placebo (Grossman et al 2006).
Daily nighttime melatonin decreased sleep systolic and diastolic blood pressure in subjects with untreated mild to moderate essential hypertension in a randomized, double-blind, placebo-controlled, crossover trial. Sixteen men (aged 36 to 68 years) were orally supplemented with melatonin 2.5 mg or placebo 1 hour before bedtime one time only (acute); this dosage was then repeated daily for 3 weeks. Twenty-four hour ambulatory blood pressure and sleep quality were evaluated for both the acute and repeated treatment period. A significant decrease in sleep systolic (6 mm Hg) and diastolic (4 mm Hg) blood pressure was demonstrated during the 3 weeks of melatonin compared to placebo (p=0.046 and p=0.020). The 1-day only treatment had no effect on systolic and diastolic blood pressures while asleep (p=0.89 and p=0.86, respectively). No significant improvement in awake systolic and diastolic blood pressure was demonstrated during the 3 weeks of melatonin compared to placebo. Repeatedly used melatonin significantly increased sleep efficiency (p=0.017) and sleep time (p=0.013) (Scheer et al, 2004).
Sedation and Anxiolysis
Contrary to earlier findings in younger subjects, a prospective, double-blind, randomized trial involving elderly subjects (>65 years old) concluded that melatonin was no more effective than placebo in reducing anxiety prior to elective surgery. Subjects were randomized to receive 10 mg melatonin (n=67) or placebo (n=71). Preoperative anxiety at 90 minutes decreased by 33% in the melatonin group and 21% in the placebo group, a nonsignificant finding. The authors point out that the differences they observed compared to earlier studies may be attributable to the higher mean age of subjects and the inclusion of males in this study, and their failure to measure plasma melatonin levels (Capuzzu et al 2006).
In a randomized, double-blind, dose-ranging study, melatonin (MLN) was as efficacious as midazolam (MZ) and more efficacious than placebo in provoking sedation and anxiolysis in preoperative surgical patients. Adult women scheduled for laparoscopic surgery were assigned to receive 1 of the following 7 sublingual premedication regimens (n=12 each group): MZ 0.05, 0.1, or 0.2 mg/kg; MLN 0.05, 0.1, or 0.2 mg/kg; or placebo. All patients were premedicated 2 hours prior to surgery; in each case, the patient was instructed to avoid swallowing the sublingual preparation of their respective study drug for 3 minutes after dosing, after which they were allowed to swallow. Sedation and anxiety were measured by a 4-point sedation scale and a visual analog scale, respectively. Both MZ and MLN induced significantly greater reductions in preoperative anxiety when compared with placebo (p<0.05), and sedation was attained at 60 and 90 minutes after dosing by significantly more patients receiving MZ and MLN than placebo (p<0.02). With the exception of patients receiving MZ 0.2 mg/kg, there were no differences between groups with regard to the level of postoperative sedation; patients receiving MZ 0.2 mg/kg experienced an increased depth of sedation compared with patients receiving MLN 0.05 and 0.1 mg/kg. Postoperative pain scores were not different between groups at any time during postanesthesia recovery (Naguib & Samarkandi, 2000).
Sleep Disorders/Delayed Sleep
A retrospective study examined the effect of long-term administration of melatonin (3 to 5 mg/day) in adolescents with delayed sleep phase syndrome. Subjects were treated for an average of 6 months as investigators evaluated changes in sleep parameters following treatment and assessed its impact on school performance and behavior. Doses of melatonin were taken 2 hours prior to bedtime in 33 subjects, and subsequently, difficulty falling asleep resolved in all of them. In this sample of patients, treatment with melatonin improved sleep-wake patterns, was accompanied by a reduction in sleep-related symptoms, and appeared to decrease the proportion of subjects who misbehaved at school (Szeinberg et al 2006). These findings are consistent with an earlier study, which found that melatonin advanced sleep onset in subjects with delayed sleep phase syndrome. When administered 5 hours before endogenous melatonin secretion, a dose of 5 mg melatonin advanced sleep onset and improved quality of life for 16 patients (Smits & Nagtegaal, 2000). One study found melatonin to be statistically and clinically superior to placebo in reducing initial insomnia, which is common among those with ADHD. In this trial, 27 children with ADHD and initial insomnia who were taking prescription stimulants were treated with sleep hygiene methods; those not responding to this treatment were randomized to treatment with melatonin (5 mg). Full remission of sleep disorders was obtained with sleep hygiene or both treatments in 71% of subjects. The response rate was significant—81% of the total sample noted clinically meaningful improvements with sleep hygiene, melatonin, or both (Weiss et al 2006). The authors of another study recommended melatonin as first-line treatment to resolve sleep disturbances in children with Sanfilippo syndrome, based on their observations of this sample population (Fraser et al 2005).
Nine patients with periodic limb movement disorder showed improvements when given melatonin for 6 weeks in an open trial. Patients were given melatonin 3 mg taken 30 minutes before bedtime, between 10 and 11 p.m. for 6 weeks. Polysomnography was performed at baseline and during the last week of treatment. A wrist actigraph was worn during 2 weeks prior to treatment and during the last 14 days of supplementation in order to measure motor activity during sleep. Seven patients subjectively reported feelings of improved daytime symptoms of fatigue and excessive sleepiness within 7 days of treatment. Polysomnographic recordings reported significant decreases in parameters of periodic limb movements (p<0.05), which were confirmed through actigraphy. The investigators hypothesized that a resynchronization of circadian rhythms enhanced REM sleep and produced the observed effects (Kunz & Bes, 2001).
Improved sleep parameters were seen when melatonin was given to schizophrenic patients with insomnia in a randomized, double-blind, placebo-controlled, crossover study. Nineteen subjects were given either controlled-release melatonin 2 mg or placebo 2 hours before bedtime for 3 weeks, with a 1 week washout between the treatments. Wrist actigraphs were used to assess sleep quality for 3 nights during the last week of each treatment period. Sleep efficiency was significantly improved in the melatonin group compared to placebo (p=0.038). There was an insignificant decrease in sleep latency and increased total sleep time with melatonin versus placebo. Twenty-seven patients entered the study, but only 19 were assessed due to lack of compliance (n=4) and technical problems with the actigraphs (n=4) (Shamir et al, 2000).
Melatonin was effective in improving the sleep of patients with major depressive disorder. This double-blind, placebo-controlled study treated 24 outpatients with fluoxetine 20 mg and either slow-release melatonin 2.5 to 10 mg or placebo nightly for 4 weeks. No other medications were used. Patients treated with melatonin showed significant improvement in sleep variables compared with controls (p=0.01). No effect of melatonin on depression symptoms was found (Dolberg et al, 1998).
Limited data suggest benefits of oral melatonin in blind adults with free-running sleep-wake rhythms (Haimov & Lavie, 1995; Cavallo, 1993; Arendt et al, 1988). Sleep problems in these patients were considered related to an inability to remain synchronized to the normal 24-hour day, due at least in part to lack of light-dark perception. Therapy with melatonin resulted in normal sleep (without early awakening and day sleeps), attributed to resynchronization of the sleep-wake activity cycle. Effective doses have been 5 mg nightly at normal bedtime.
Ten children with neurologic and developmental disorders were given 3 mg melatonin 1 to 2 hours before bedtime and monitored for a mean of 7.5 months. In eight (80%) of the children, the average number of hours of sleep, average number of nighttime awakenings, average number of nights with early morning arousal, and average number of nights with delayed sleep onset greatly improved after treatment. Use of melatonin also improved daytime activities and behaviors (Jan, 2000). Another study examining 5 severely psychomotor retarded children showed improved sleep patterns after 4 weeks of treatment with 3 mg melatonin given every day at 6:30 p.m. (Pillar et al, 2000).
In a study of 100 pediatric patients (ages 3 months to 21 years) with chronic sleep disorders, 54% were visually impaired and 85% had multiple neurodevelopmental disabilities (including visual loss, deafness, blindness, mental retardation, cerebral palsy, epilepsy, chromosomal abnormalities, head injuries, degenerative central nervous system disorders, autism, and brain tumors). Of the 15 patients not having the above conditions, diagnoses were attention deficit hyperactive disorder, anxiety, bowel disorders, and nocturnal seizures. Treatment with oral melatonin doses of 2.5 to 10 mg resulted in improved sleep in 82% of these patients. Factors involved in non-response were recurrent pain, malfunction or absence of the suprachiasmatic nucleus, noisy sleep environments, psychological reasons for delayed sleep onset, multiple medications, and organically driven behavior. In 2 children with pineal glands destroyed by brain tumor or trauma, oral melatonin doses of 25 mg were required to produce a beneficial improvement in sleep quality. A secondary benefit was that improved patient sleep allowed caretakers to also have improved sleep (Jan & O'Donnell, 1996).
Patients reported more subjective improvement in tinnitus with melatonin than with placebo in a randomized, crossover study. Twenty-three subjects were given melatonin 3 mg or placebo, to be taken 1 to 2 hours before retiring, for 30 days. After a 7-day washout, patients took the alternate treatment for 30 days. Average improvements in Tinnitus Handicap Inventory (THI) scores were significant and were the same with melatonin and placebo. However, on a follow-up questionnaire, 39% of patients reported an overall improvement with melatonin, whereas only 17% reported improvement with placebo. Among those reporting difficulty sleeping due to their tinnitus, 46.7% had overall improvement with melatonin and 20% with placebo (p=0.04). Eight of 16 patients with bilateral tinnitus reported improvement with melatonin and only 3 of the 16 had improvement with placebo. Overall, 35% of subjects had a decrease in loudness of their tinnitus with melatonin, compared with 13% with placebo. The only side effect reported in the study was bad dreams, which was equally distributed between the melatonin and placebo trials (Rosenberg et al, 1998).
Indications & Usage
Melatonin is used as a sleep aid, in the treatment of a variety of solid tumors (in combination with interleukin-2), and can be used to improve thrombocytopenia and other toxicities induced by cancer chemotherapy and from other conditions. Melatonin is also used for cluster headaches and tinnitus.
Precautions & Adverse Reactions
Adverse effects of exogenous melatonin have generally been minimal. Drowsiness, fatigue, headache, confusion, gastrointestinal complaints, and reduced body temperature have been reported. Melatonin has exacerbated dysphoria in depressed patients and has caused mood swings. It has caused depressive symptoms and fever when used with interleukin-2 in cancer therapy. Rarely, tachycardia, seizures, acute psychotic reactions, autoimmune hepatitis, and pruritus have been reported.
Concomitant use may result in an increased risk of bleeding. Clinical Management: Avoid concomitant use of melatonin and warfarin. If both agents are taken together, monitor prothrombin time, INR, and signs and symptoms of excessive bleeding frequently. Only adjust the warfarin dose if the patient takes a consistent dosage of melatonin with a consistent and standardized brand.
Concomitant use may result in increased central nervous system depression. Clinical Management: Monitor patients taking fluvoxamine with melatonin supplementation for changes in sleep patterns and signs of excessive central nervous system depression. Downward titration of melatonin dosages may be required during concomitant administration with fluvoxamine.
Concomitant use may result in increased blood pressure. Clinical Management: Close monitoring of blood pressure is advised with appropriate dose adjustment of nifedipine or withdrawal of melatonin.
Mode of Administration
intramuscular, intravenous, oral, oral transmucosal, transdermal
- Cancer as combination therapy: 40 or 50 mg oral tablets given once daily at night, initiated 7 days prior to interleukin-2 and continued throughout the cycle
- Cancer as single agent therapy: Melatonin 20 mg intramuscularly daily for 2 months (induction phase), followed by oral doses of 10 mg daily until progression, has been given for the treatment of solid tumors
- Chronic insomnia: 1 to 10 mg orally daily
- Delayed sleep phase syndrome: 5 mg orally daily
- Jet lag: 5 mg orally daily for 3 days prior to departure, then 5 mg for 4 additional days
- Normalization of nocturnal levels: constant intravenous daytime infusion of melatonin 4 mcg/hour for 5 hours (0.1 and 0.3 mg oral tablets have also been used)
- Sleep disorders: 5 mg at the usual bedtime
- Congenital sleep disorder: 2.5 mg oral tablet
- Neurological disability: 0.5 to 10 mg oral tablet at bedtime
- Sleep-wake cycle disorder (controlled release tablets, oral: An average of 5.7 mg was used in children ages 4 to 21 years