Study at a glance
How were vitamin D levels, sleep and melatonin studied?
Researchers studied vitamin D levels, sleep and a urine marker of overnight melatonin, a sleep-related hormone, in 79 apparently healthy adults aged 18 to 45. Participants wore a wrist movement monitor for about two weeks. The device estimated sleep from movement; it did not record brain sleep stages. No vitamin D treatment was assigned.
Were higher vitamin D levels linked with longer or better sleep?
Before allowing for other differences between people, higher vitamin D was linked to several more favourable sleep measurements estimated by the device. After accounting for age, sex, race and body fat, only longer sleep duration remained statistically significant. Other analyses used different sets of adjustments rather than accounting for every factor together. The results therefore did not show that all sleep measures were consistently linked to vitamin D.
Did melatonin explain the link between vitamin D and sleep?
The single-night urine marker was not significantly related to vitamin D or the sleep measures. The researchers therefore did not carry out a formal analysis testing whether melatonin explained the association. This finding did not rule out every role for melatonin in sleep. Self-reported sleep duration and overall sleep-quality scores also did not show significant links with vitamin D.
Does the study show that vitamin D can treat sleep problems?
The study observed people at one period and could not show cause and effect. Sleep habits, outdoor time, the amounts of fat and other tissue in the body, or other influences could affect both vitamin D and sleep. One urine collection could not describe the full timing of melatonin release. The selected healthy group excluded shift workers and people with clinical sleep disorders. These findings do not establish insomnia treatment, an ideal vitamin D level or the best time to take supplements.
Read the full article below ↓Higher blood vitamin D was associated with several device-estimated sleep measures in a study of 79 adults, but the pattern changed substantially after adjustment. Only sleep duration remained significant in the model including age, sex, race and body fat. An overnight urinary melatonin measure did not support the proposed melatonin explanation. For the shorter account, see Vitamin D Status, Actigraphy and Overnight Melatonin: Study Summary.
This is a VD3FL analysis of the complete principal article, its three tables and main figure. It reports a cross-sectional study, not a vitamin D treatment trial. Findings and research definitions are not personal treatment targets.
What question did the study ask?
D’Agata and colleagues investigated whether serum 25-hydroxyvitamin D was related to habitual sleep measured with a wrist device, and whether overnight melatonin production might help explain any association. The study combined these measurements in the same participants, making it more informative than a comparison based solely on a single sleep questionnaire.
Its design was nevertheless cross-sectional. Vitamin D status, sleep patterns and the hormone proxy were observed around the same period; vitamin D supplementation was not randomly assigned. The findings add to Vitamin D and Sleep by examining measurement-specific associations, rather than showing that raising vitamin D changes sleep.
Who took part?
The analysis included 79 apparently healthy adults aged 18-45, with a mean age of 29. There were 52 women and 27 men. Participants came from the University of Delaware and the surrounding Newark area; 80% had at least a bachelor’s degree, and 53% were non-Hispanic White.
Shift workers, people with clinical sleep disorders or depression, smokers, and people with relevant major illness or medication use were excluded. This selection helped reduce some sources of variation, but limits transfer of the results to insomnia patients, shift workers or people with more complex health conditions.
Data came from two protocols using the same measurement procedures. One contributed baseline observations from a registered trial. Baseline observations do not become an intervention comparison merely because a parent project is a trial. In the present paper, no assigned dose, treatment duration or morning-versus-evening dosing comparison was analysed.
What the wrist device measured
Participants wore an Actiwatch Spectrum Plus for 14 days, with at least ten valid days required and an average of 13 achieved. Sleep diaries helped identify the relevant intervals. The device classified movement-based sleep and waking in 30-second epochs using the manufacturer’s medium threshold.
The outcomes included sleep duration, efficiency, variation in duration, variation in sleep onset and midpoint, and the sleep regularity index. Efficiency describes the proportion of the relevant sleep interval estimated as sleep. A variability measure asks how much timing or duration changes across nights; it is not simply the average amount of sleep.
The sleep regularity index compares estimated sleep or waking at times 24 hours apart. It therefore captures consistency in a way that differs from average duration. These distinctions connect the paper with Vitamin D and Sleep: Wake Regulation. Someone can sleep longer without becoming more regular, or maintain a regular schedule while sleeping relatively briefly.
Actigraphy estimates sleep from movement. It does not directly record EEG sleep stages or the phase of the internal circadian clock. The study’s mean device-estimated duration was 6.8 hours and mean efficiency 84%. Self-reported duration averaged 7.4 hours, illustrating that the two methods were measuring related but nonidentical accounts of sleep.
Vitamin D status and the melatonin proxy
A fasting morning blood sample was collected after the final monitoring night. Serum 25-hydroxyvitamin D was measured in triplicate by an immunoassay. The mean concentration was 34.5 ng/mL. The paper divided participants into study-defined status categories, with 11 below 20 ng/mL, 23 between 20 and 29.9, and 45 at or above 30.
Those category boundaries describe the study; they do not establish a universal treatment threshold or an optimal sleep target. Vitamin D Testing: 25-OH-D vs 1,25-OH₂-D explains why the identity of the measured metabolite matters. This was a nutritional-status marker, not a measurement of active vitamin D inside the brain.
For melatonin, participants pooled all overnight urine and the first morning void. The analysis used 6-sulfatoxymelatonin, abbreviated aMT6s, relative to urinary creatinine. This is a proxy for overnight melatonin output. Despite a section heading referring to clearance, the analysis did not independently measure a creatinine clearance rate.
Two urine samples lay outside assay detection limits, leaving 77 participants for that measurement. One pooled night cannot establish a full melatonin secretion curve, dim-light melatonin onset or hormone patterns throughout the two-week device period. The connection to Vitamin D and Neuroendocrine Integration therefore concerns a limited endocrine measurement, not comprehensive hormone coordination.
At-home light exposure was not constrained or measured across the full day. Light, behaviour and season could influence the relationship between these variables, while the study supplied no direct assessment of circadian phase. That distinction is central to interpreting the broader subject of Vitamin D and Circadian Biology.
The unadjusted associations
Before covariate adjustment, higher vitamin D status was significantly associated with all six device outcomes. For each 10 ng/mL difference between people, the reported associations corresponded to about 12 minutes longer sleep, 1.7 percentage points higher efficiency, five minutes less duration variability, 9.5 minutes less onset variability, 5.1 minutes less midpoint variability and 1.3 points higher sleep regularity.
These are cross-person statistical associations, not measured gains after taking a supplement. In particular, percentage points in efficiency and points on the regularity index should not be described as equivalent relative percentage improvements. A person moving their blood concentration by 10 ng/mL was not the experiment performed.
The adjusted duration coefficient was smaller than the unadjusted one. The rounded Model 2 estimate of .01 hours per ng/mL corresponds to approximately six minutes per 10 ng/mL, rather than the unadjusted twelve-minute association. This arithmetic illustrates attenuation; it should not be treated as a precise expected benefit for an individual, especially because printed coefficients and intervals were rounded.
Even an association that survives adjustment is conditional on the covariates selected, their measurement quality and the model form. A wrist device can improve on recall for some purposes without eliminating confounding in the exposure-outcome relationship. Objective measurement of one variable is not equivalent to random assignment of another.
The six outcomes were correlated features of the same people’s sleep. Finding several associations therefore does not mean six independent studies confirmed the result. The paper used a Benjamini-Hochberg multiple-testing approach within its models, with a reported threshold of P≤.0375.
Adjustment changed the conclusion
The models were alternatives, rather than a single progressively adjusted sequence. Model 2 included age, sex, race and percentage body fat. Only duration remained significant there, with P=.036. The other five device outcomes did not meet the study’s significance criterion in that model.
Body composition is one possible common influence on status and sleep. Vitamin D and Body Fat provides the broader physiological topic, but the disappearance of an association after adjustment does not prove that body fat alone explains it. Several covariates were included together, and the study was not designed to isolate a causal role for each.
Model 3 instead considered season, caffeine and education, without retaining the demographic and body-fat covariates. Duration, efficiency and variability of onset and midpoint remained significant. Duration variability and regularity did not. Vitamin D and Season is relevant because seasonal sampling can change the context of a vitamin D measurement.
Model 4 considered physical activity in the smaller subset with available activity data, 68 people, and did not simultaneously include all the other covariates. All six outcomes were significant there. It would be misleading to call those associations robust to every listed confounder at once.
The table contains apparent reporting errors. The printed interval for Model 2’s regularity estimate does not agree with its point estimate or P value. Another coefficient has a scale inconsistency. This analysis relies on the authors’ explicitly stated model-level significance pattern and does not silently repair those intervals or use them to calculate individual predictions.
The melatonin hypothesis was not supported
Vitamin D status was not significantly related to the aMT6s-to-creatinine ratio, P=.65. The urine ratio was also not significantly associated with the sleep outcomes, all P values above .31. Under the authors’ stated approach, these absent preliminary associations meant that a formal mediation analysis was not performed.
The precise conclusion is that the measured overnight melatonin explanation was not supported in this dataset. It is stronger than the evidence to claim that a formal analysis proved zero mediation, or that melatonin has no general role in sleep. The measurement and sampling limits leave narrower possibilities unresolved.
Nor did the study show significant associations with global Pittsburgh Sleep Quality Index scores or self-reported duration, with P=.81 and P=.07 respectively. Device measures and personal reports should retain their different meanings. The paper does not establish that one method is universally correct and the other dispensable.
Why this is not a supplementation result
Thirteen people reported supplement use. They had higher vitamin D concentrations on average, but did not show significant differences in the urine ratio or sleep metrics. This small, self-selected comparison had unknown doses and frequencies; it cannot establish either a treatment benefit or proof that supplementation has no effect.
Reverse causation and common causes remain possible. Sleep patterns may influence time outdoors, activity or other behaviours related to status. A single blood sample cannot reconstruct long-term exposure. Vitamin D Response helps distinguish a measured status difference from an intervention-induced biological response.
The authors disclosed research support from the American Heart Association, University of Delaware and NIH/NIGMS, with no financial or nonfinancial conflicts declared. The analysis did not establish clinical insomnia improvement, cardiovascular benefit or an optimal dosing time.
Conclusion
Higher vitamin D status tracked several device-estimated sleep measures before adjustment, but only duration remained significant in the demographic and body-fat model. The single-night urine marker did not support the proposed melatonin explanation, and formal mediation was not performed. The study identifies associations worth investigating while leaving treatment effects and circadian mechanisms unresolved.
Scientific sources
D’Agata MN, Hoopes EK, Keiser T, Patterson F, Brewer BC, Witman MA. Serum 25-hydroxyvitamin D concentration is associated with device-estimated sleep metrics in healthy young and early middle-aged adults. Sleep Advances. 2025;6(4):zpaf077. DOI 10.1093/sleepadvances/zpaf077; PMID 41341996. Readable full text is available at PubMed Central.
Related reading
Controlling the Progression of Curvature in Children and Adolescent Idiopathic Scoliosis Following the Administration of Melatonin, Calcium, and Vitamin D
In this perspective, randomized, case-control, interventional study, the impact of using melatonin, calcium, and vitamin D, respectively, on idiopathic scoliosis patients was analyzed.
Original paper ↗Read more →Evaluating the antioxidant and anti-inflammatory effect of melatonin in pediatric hemodialysis patients: a randomized, placebo-controlled trial
This prospective, block-randomized, double-blinded, placebo-controlled study aimed at assessing the effect of 5 mg MLT on oxidative stress and inflammation in pediatric hemodialysis patients.
Original paper ↗Read more →Serotonin, melatonin and their precursors and metabolites and vitamin D3 derivatives in honey
We are commenting recent discoveries on the presence of L-DOPA, dopamine, 5-hydroxytryptophan, tryptamine, serotonin, N-acetylserotonin, melatonin, 2-hydroxymelatonin, AFMK and AMK in honey.
Original paper ↗Read more →Serum 25-hydroxyvitamin D concentration is associated with device-estimated sleep metrics in healthy young and early middle-aged adults
Participants (n = 79) completed 14 days of wrist actigraphy.
Original paper ↗Read more →The effect of Folic Acid, B12, D, and E Vitamins and Melatonin levels in the follicular fluid taken by the Intracytoplasmic Sperm Injection method on pregnancy
We examined the effect of folic acid, B12, D, and E vitamins and melatonin values in follicular fluid on pregnancy in women with low, normal, and high ovarian reserves who underwent ICSI.
Original paper ↗Read more →A pilot study investigating the effect of serum 25-hydroxyvitamin D on device-estimated sleep metrics in young black women: a single-arm open-label clinical trial
Black women (BLW) experience a higher prevalence of vitamin D deficiency and short and disturbed sleep compared with other races.
Original paper ↗Read more →The Association between Sleep Duration and 25-Hydroxyvitamin D Concentration with Obesity in an Elderly Korean Population: A Cross-Sectional Study
Studies have recently reported an association between sleep duration and obesity in some individuals.
Original paper ↗Read more →In search of vitamin sleep: cross-sectional analysis of vitamin and mineral intake and sleep health outcomes in adolescents
We examined how adolescent micronutrient intake is related to sleep.
Original paper ↗Read more →In search of vitamin sleep: cross-sectional analysis of vitamin and mineral intake and sleep health outcomes in adolescents
We examined how adolescent micronutrient intake is related to sleep.
Original paper ↗Read more →Vitamin D Status, Actigraphy and Overnight Melatonin: Study Summary
Summary of a 79-adult observational study on vitamin D, wrist-device sleep estimates and overnight melatonin, explaining the effects of statistical adjustment.
Read more →A comparative cross-sectional study on sleep quality in patients with a history of differentiated thyroid carcinoma and its correlation with quality of life
To evaluate sleep quality in differentiated thyroid carcinoma (DTC) patients and correlate sleep disturbances with quality of life (QoL).
Original paper ↗Read more →Dietary Magnesium Intake Is Associated With Self‐Reported Short Sleep Duration but Not Self‐Reported Sleep Disorder
The study dataset includes participants from five cycles (2009 to 2018) of NHANES.
Original paper ↗Read more →DXA-Derived Total Body Fat Percentage and Serum 25(OH)D in Overweight and Obese Children: A Cross-Sectional Study
Participants with and without DXA were similar in age, sex, BMI z-score, and serum 25(OH)D, but blood sampling season differed (p < 0.001).
Original paper ↗Read more →Relationship between Sleep Duration, Sun Exposure, and Serum 25-Hydroxyvitamin D Status: A Cross-sectional Study
Normal-range sleep duration is an important factor for general health and metabolism, and insufficient or excessive sleep is associated with chronic metabolic disease.
Original paper ↗Read more →Sleep duration is associated with vitamin D deficiency in older women living in Macao, China: A pilot cross-sectional study
Chinese women are known to have both a high prevalence of metabolic syndrome (MetS) and vitamin D deficiency (serum 25-hydroxyvitamin D (25OHD).
Original paper ↗Read more →The association of Malnutrition-Inflammation Score with sleep quality and mental health in hemodialysis patients: a multicenter cross-sectional study
Therefore, the study aimed to investigate the association between the Malnutrition-Inflammation Score (MIS) and various health parameters, including sleep quality and mental health in hemodialysis patients.
Original paper ↗Read more →Latitude, Vitamin D, Melatonin, and Gut Microbiota Act in Concert to Initiate Multiple Sclerosis: A New Mechanistic Pathway
Multiple sclerosis (MS) is an inflammatory demyelinating disease of the central nervous system (CNS).
Original paper ↗Read more →Predictors of vitamin D deficiency and quality of life in obese patients with obstructive sleep apnea
In this cross-sectional study, we aimed to determine the vitamin D profile, evaluate the impact of vitamin D level on quality of life (QoL), and identify factors associated with hypovitaminosis D in obese patients with OSA.
Original paper ↗Read more →Unveiling the benefits of Vitamin D3 with SGLT-2 inhibitors for hypertensive obese obstructive sleep apnea patients
As vitamin D3 and SGLT2i act through different pathways with similar mechanisms in improving cardio-metabolic health, this study aimed to investigate the synergistic effects of this combination therapy in improving these parameters and QoL in hypertensive obese OSA patients.
Original paper ↗Read more →Vitamin D and Body Fat
Learn how body fat influences vitamin D storage, availability, and distribution, and why individuals with different body compositions may show different vitamin D levels and responses.
Read more →Vitamin D and Circadian Biology
Explains how vitamin D relates to circadian biology, including links with light exposure, hormone timing, metabolism, immune rhythms, and sleep: wake regulation, without making treatment claims.
Read more →Original sources
Serum 25-hydroxyvitamin D concentration is associated with device-estimated sleep metrics in healthy young and early middle-aged adults
Participants (n = 79) completed 14 days of wrist actigraphy.
Publication details
D’Agata MN, Hoopes EK, Keiser T, Patterson F, Brewer BC, Witman MA. Serum 25-hydroxyvitamin D concentration is associated with device-estimated sleep metrics in healthy young and early middle-aged adults. Sleep Advances. 2025;6(4):zpaf077. DOI 10.1093/sleepadvances/zpaf077; PMID 41341996. Readable full text is available at PubMed Central.