Predictions of melatonin suppression during the early biological night and their implications for residential light exposures prior to sleeping

The magnitude of nocturnal melatonin suppression depends upon the spectrum, amount, and duration of light exposure. The functional relationship between melatonin suppression and the light spectrum and amount have been previously described. Only one duration-dependent parameter was needed to extend this functional relationship to predict nocturnal melatonin suppression during the early biological night from a variety of published studies. Those predictions suggest that ambient lighting commonly found in North American homes will not suppress melatonin for durations up to 3 h, whereas extended use of self-luminous displays in the home prior to sleep can.


Results -Validation
The criteria for inclusion of studies in the validation exercise are listed in Supplementary Table  S2. Brief overviews of the studies included in the validation exercise are provided in Supplementary  Table S3. Some of these studies calculated melatonin suppression (e.g., area under the curve, or AUC) using a different method than that employed by Nagare, et al. 1 . Moreover, the data points may not have been available in the public domain, so the plots relating the light stimulus to melatonin concentrations or suppression were digitized using a freely available online tool (WebPlotDigitizer ver. 4.2, Ankit Rohatgi, San Francisco, California, USA) and point estimates for different exposure durations were determined.

Supplementary
From these estimates, and to be consistent with Nagare, et al. 1 , melatonin concentrations at each sampling time point were normalized to the melatonin value for the first sample taken under that condition. Melatonin suppression was then determined by comparing the normalized melatonin levels collected during the dim light control condition to the normalized levels collected at the corresponding time for each lighting intervention using the following formula: where Mn is the normalized melatonin concentration at a given time point following a light exposure and Md is the normalized melatonin concentration at the same time point during the control night. Cajochen, et al. 4 Melatonin suppression was determined using 10 male subjects exposed to 2 h (21:30 -23:30) of narrowband (460 nm or 555 nm) light, both at approximately 2.8 x 10 13 photons cm -2 s -1 Some of these data could not be used because melatonin concentrations did not rise in the control condition after 1 h.

Study Experimental design
Only the data made public by the authors for 2 representative subjects exposed to a dose of 12.85 log photons cm -2 s -1 could be used. Data for the 555 nm condition could not be included in the validation exercise because calculated 1-h melatonin suppression was zero. Kozaki, et al. 8 Melatonin suppression was determined using 15 male subjects exposed to 1.5-h (01:00 -02:30) of photon density matched flickering and non-flickering blue light (~ 14 lx) Kraneburg, et al. 9 Melatonin suppression was determined using 16 healthy subjects exposed for 4-h ( 22:00 -02:00) to 7 spectrally distinct polychromatic white light sources (CCTs: 1600 K -14,000 K) delivering 200 lx at the eye Some of these data could not be used because melatonin concentrations did not rise in the control condition after 2.5 h. Calculated 1-h melatonin suppression values for 2 light sources from Kraneburg et al. (2017) (CCTs: 1600 K, 1950 K) were both zero, and hence, no data points could be included for these spectra.
Nagare, et al. 10 Melatonin suppression was determined using 12 adults and 12 adolescents exposed to 4-h (23:00 -03:00) of polychromatic white light (2700 K, 5600 K) delivering a CS of 0.25 at the eye Rahman, et al. 11 Melatonin suppression was determined using 12 adult subjects exposed to 12-h (20:00 -08:00) of polychromatic light after completely filtering light at wavelengths < 480 nm, wavelengths < 460 nm, or partially filtering light at wavelengths < 480 nm Data for one condition (100% filter < 480 nm) could not be included in the validation exercise because calculated 1-h melatonin suppression was zero.
Souman, et al. 12 Melatonin suppression was determined using 15 healthy subjects exposed to 3-h (starting 2 h prior to habitual bedtime) of white light (2 sources with high or low radiant power in μW cm -2 between 450 -500 nm) matched for photopic illuminance of 175 lx Data for one condition (low power between 450 -500 nm) could not be included in the validation exercise as calculated 1-h melatonin suppression was zero.
Van de Werken, et al. 13 Melatonin suppression was determined using 33 male subjects exposed to 8-h (23:00 -07:00) of short-wavelength attenuated polychromatic white light (193 lx), or fullspectrum light (256 lx) Data for one condition (short-wavelength attenuated light) could not be included in the validation exercise because calculated 1-h melatonin suppression was zero.
West, et al. 14 Melatonin suppression was determined using 8 adult subjects exposed to 1.