Biologically aware lighting for newborn intensive care

We designed and implemented a novel neonatal intensive care (NICU) lighting system to support the current understanding of daylight-coupled physiology. We created a system that generates wavelengths corresponding to the known blue and violet activation spectra of non-visual opsins. These are known to mediate energy management and related physiologic activity. Light produced by the system spans the visible spectrum, including violet wavelengths that are blocked by modern glazing and not emitted by standard LED fixtures. System features include automated light and dark phases that mimic dawn/dusk. The system also matches length of day seasonality. Spectral composition can be varied to support translational research protocols. Implementation required a comprehensive strategy to inform bedside providers about the value and use of the lighting system. Full-spectrum lighting for the NICU is feasible and will inform the optimization of the NICU environment of care to support optimal neonatal growth and development.


INTRODUCTION
The most dynamic phases of growth, metabolic activity, and organ development during the human lifespan begin with the fetal period and continue through the early years of extrauterine life.There is much evidence that this crucial period of development is profoundly impacted by a myriad of environmental stimuli such as sound and light.For example, the plasticity of the auditory cortex in preterm infants is preferentially responsive to intrauterine maternal sounds [1].Another example emerges from studies of circadian stimuli in neonates.The onset of cyclical patterns is detectable during the early second trimester and becomes more intense as term gestation approaches [2].Clinical studies of basic cycled lighting environments point to a positive impact on weight gain, growth, and length of stay [3][4][5] when compared with conventional lighting.
Environmental stimuli experienced by hospitalized infants in the NICU are very different from those within the intrauterine setting.When a birth is preterm, these contrasts are exacerbated by the early termination of environmental inputs provided by the mother.The NICU setting confronts the newborn with a litany of unnatural stimuli ranging from equipment alarms to artificial lighting.
As we consider the impact of the environment on early growth and development, we must include a growing body of new knowledge regarding the biology of light sensing.Beyond cycled (light/dark) lighting, environmental light exerts specific biological effects through light-sensitive G-protein coupled receptors, known collectively as opsins (Table 1).In humans, these respond to wavelength ranges within the 380-700 nm range of the visible spectrum [6][7][8].Of particular interest, three opsins detect light but do not directly contribute to vision.OPN4 (melanopsin) has generated substantial investigation following its confirmed functional role in circadian entrainment influencing sleep/wake cycles [9].The functions of two additional non-visual opsins, OPN3 (encephalopsin) and OPN5 (neuropsin) are now understood to impact fundamental physiologic and metabolic processes including temperature regulation, energy management, and growth [10,11].Violet stimulation of OPN5 promotes anabolic pathways and blue stimulation of OPN3 activates catabolic activity.Under conditions of conventional artificial lighting, OPN5 is not activated.Modern glazing products block violet wavelengths and cannot compensate for the limited spectral transmission of artificial light sources.Therefore, it is plausible that the chronic absence of OPN5 activation may blunt anabolic activity needed for optimal growth.These findings have led to the recognition that mammalian physiology is intimately coupled to the day-night cycle and requires interaction with wavelengths not produced by standard artificial lighting sources.As knowledge about these evolutionarily conserved nonvisual opsins expands, it is very likely that their activation is necessary for optimal NICU outcomes [12].To address this idea, we generated a NICU lighting system capable of satisfying the needs of daylight-coupled physiology (DCP).
The construction of a new Critical Care Building (CCB) at Cincinnati Children's Hospital Medical Center (CCHMC) offered a unique opportunity to incorporate DCP lighting into clinical care settings.We chose the level IV NICU as the installation site in the CCB given the relatively long average length of stay (approximately 35 days) and the very dynamic growth and development intrinsic to this patient population.Here we detail the design and deployment of that lighting system and how we will use this system to advance understanding and improve outcomes for the NICU patient population.

Design specifications
We sought to create a NICU lighting system that satisfies established design specifications presented in detail by White et al. [13].We also incorporated relevant general design and architectural standards such as those published by the Illuminating Engineering Society (IES).We envisioned this lighting system not only as a standard patient room lighting solution but also as a device to support circadian and opsin-based translational research.Within this conceptual framework, we defined four fundamental performance characteristics: 1. Automated diurnal light cycling to reflect day-night cycles, 2. Generation of light wavelengths for activation of visual and nonvisual opsins (Table 1), 3. Dynamic spectral variation corresponding to daily sunrise-middaysunset progression, 4. Length of day seasonal variation.
Our concept required a control system that supported the functionality noted above and allowed for variation in spectral power distribution, lightdark cycles, and seasonality.We also required a lighting control system that could be easily managed by bedside care providers with minimal disruption to patient care activities.To support the functionality required for NICU patient care operations, the luminaire generates high-quality broad-spectrum white light at a 3500 K color temperature consistent with standard lighting specifications at Cincinnati Children's.The lighting generates a cyanosis observation index threshold of <3.3 to support accurate patient observations based on Australia/New Zealand standard AS/NZS 1680.2.5:2018.

Daylight and artificial light spectral analysis
Daylight, defined as the sum of direct and indirect solar irradiance, was collected through a calibrated spectrometer with a peak sensitivity range of 350-800 nm (STS-VIS, Ocean Insight, Orlando, FL).We mounted the spectrometer on a mast located 25 feet above the roofline of the 14-story Cincinnati Children's Clinical Sciences Pavilion located about 500 feet from the CCB.The location allows for 97.5% coverage from horizon to zenith in all directions.Minimal shadowing (2.5%) is due to an air terminal for lightning protection located a few feet north of the spectrometer detector.The detector is a 1024-pixel CMOS device equipped with a cosine corrector housed in a weather protection dome and connected to the spectrometer via fiberoptic cable.Data was collected every 10 min and downloaded to data acquisition software.In addition, we obtained spectral measurements of the standard ambient lighting found in the former CCHMC NICU ("old NICU," closed in November 2021) and the CCB NICU ("new NICU").These measurements employed a custom-built spectrometer designed and manufactured by Ocean Insight.This device contained four STS-VIS spectrometers to capture irradiance from a variety of locations within the NICU environment including open cribs, radiant warmers, and isolettes (with and without coverings).Additional details regarding the lighting detection systems are provided in Cao et al. (submitted for publication).

Luminaire manufacture and installation
Once all spectral and luminaire design parameters were defined, we assembled a multidisciplinary manufacturing team (BIOS LLC, Acuity Brands, DGLogik, Pivotal Lighting Design) to create working luminaire prototypes and develop a control interface that could support various user needs ranging from bedside care providers to researchers.Two years prior to CCB patient occupancy we organized a multidisciplinary team (the "NICU Lighting Group") to oversee the implementation and installation of the lighting system.Members included the Lead Unit Educator for the NICU, a bedside RN, an Advanced Practice Nurse, a Neonatologist, and representatives of the design and construction teams.The team met monthly to review design and manufacturing progress.Their work included the evaluation of working prototypes and active participation in the design of the control interface and lighting software.Each working wall-mounted prototype was evaluated in-person by the design and research teams in a full-scale model patient room.Family representatives also participated in prototype evaluation as a component of the CCB design process.Evaluation parameters included extensive measurements of spectra, functionality of the control interface, and qualitative assessment of light esthetics.Patient room installation required coordination with the CCB construction manager, subcontractors, and the Cincinnati Children's information technology, physical plant, and clinical engineering departments.The prototype evaluation phase documented the need for a portable version of the lighting system to address potential areas of limited lighting coverage within certain patient settings and the possibility of research settings outside the CCHMC NICU.The portable fixture meets the same lighting specifications as the wall-mounted fixture.The manufacturing team generated luminaires for installation into each of the 56 NICU patient rooms located on the fourth floor of the new CCB.The NICU also includes an additional 36 patient rooms located on the first floor of the CCB with standard light emitting diode (LED) lighting.All lighting was active on November 7, 2021, the time of initial patient occupancy.

Daylight and NICU spectral analysis
Our outdoor spectrometer measurements detected a comprehensive span of wavelengths between 350-800 nm.As expected, the

Luminaire design and performance
The spectral lighting system was installed in 56 patient rooms in the CCB and complements additional standard LED fixtures that are incorporated in all CCB patient rooms.The spectral lighting system luminaire includes a 6 LED light engine capable of generating all wavelengths present in natural daylight with an emphasis on visual and non-visual opsin activation.Separate controls are provided for each LED channel allowing for variation in the spectral composition to emulate the sunrise-midday-sunset progression.The spectral output of each NICU luminaire can be modulated to support clinical research protocols to interrogate opsin function.For example, wavelengths corresponding to OPN5 activation can be omitted from the standard daytime progression to mimic standard artificial lighting conditions.Similarly, the length of day can follow seasonal variation or remain static.The system software supports a lighting condition archive for research studies.Researchers can assign lighting protocols for each patient room by remote access.The in-room control interface allows a bedside provider to assume control of the lighting system at any time if necessary for patient care.The NICU Lighting Group determined that the spectral lighting system would be inactivated for clinical conditions requiring a low-light environment such as pulmonary hypertension, opioid dependence or during ECMO support.Bedside care providers, physicians, and families could also request a low-light environment at their discretion.The control system also allowed temporary inactivation of the lighting system for diagnostic procedures such as transillumination and bedside ultrasound examinations (Fig. 2).

Clinical practice
We put considerable effort into pre-move in-service education and training.Nonetheless, the transition from the old NICU to the new CCB NICU brought dramatic changes in patient room configuration.In addition to the spectral lighting system, the new NICU rooms also included LED fixtures that generated an order of magnitude increase in light levels compared to the old NICU environment.Consequently, providers found the lighting environment esthetics bright in contrast to the old NICU promoting some providers to inactivate the lighting system.Providers also raised concerns regarding patient sleep, exacerbation of certain pathophysiologic conditions such as pulmonary hypertension, and family acceptance.We summarize our approaches to engage and support staff from design through CCB occupancy in Table 2.The NICU Lighting Group collected feedback from patient families and bedside care staff through a combination of weekly rounds, presentations at monthly staff meetings, and ad hoc in-service education.We found approaching staff for brief (10 min) discussions at the bedside with timing based on their individual workflow led to more engagement and interest than formally scheduled in-service sessions.Feedback from these encounters as well as post-occupancy survey information was reviewed and translated into refinement of the system during the first 8 months of CCB residence.We incorporated our findings into a new NICU clinical practice guideline issued in its final form 6 months after the CCB patient move.The guideline covers recommended light intensity for patients under 32 weeks corrected gestation and the elimination of isolette coverings for all patients.Low light for napping is implemented on an ad hoc basis by the bedside provider.Patients may be clothed as deemed appropriate by staff and families.Ongoing continuing education, staff presentations, and informal rounding supported the adoption of the guideline and increased utilization of spectral lighting (Fig. 3).

Clinical studies
The design of the NICU lighting system supports research in a realworld clinical setting.Our current understanding of opsin biology points to their importance for energy management and therefore their potential impact on growth and development.While the past 25 years have seen a dramatic improvement in our ability to provide nutritional support for the NICU population, many important questions remain outstanding.For example, protein and energy intake predict brain growth.However, optimization of neurodevelopmental outcomes through nutritional management remains a challenge [14].Appropriate stimulation of OPN3 and OPN5 may impact the regulation of anabolic and catabolic activity and now can be interrogated through our NICU lighting system.Similarly, the development of normal sleep-wake patterns is likely impacted by access to circadian signals that go beyond simple light-dark cycles.

DISCUSSION
We present the first lighting system to be designed to accommodate the new knowledge of daylight-coupled physiology.This extends cycled lighting to incorporate spectral elements of the visible spectrum that activate non-visual opsins.Previous designs focused only on light-dark cycles or daytime color temperature modifications that do not include these wavelengths [15,16].To our knowledge, this is the only cycled lighting system that generates wavelengths capable of OPN5 activation.This allows careful interrogation of the activity of OPN5 in a clinical setting.While the NICU is a compelling place for this initial phase of translational investigation, many opportunities exist in other Fig. 2 A sample spectral distribution recipe for light generated in the Critical Care Building NICU.The lighting system luminaire is horizontally mounted above the patient headwall (inset).The intensity and spectral distribution from dawn through dusk is controlled by custom software based on daylight spectral data obtained from the rooftop spectrometer (see Fig. 1).The bars to the right of each panel depict the research control interface which can be programmed to generate variable spectra throughout the daylight period for clinical studies.
locations such as long-term behavioral health facilities and other sites for the management chronic conditions.A growing body of directly implicates OPN5 and violet light exposure to the progression of myopia and retinopathy of prematurity [17,18].There is also substantial interest in the impact of the cycled lighting environment on providers and families.The ability to move from generic light-dark cycling to dynamic full-spectrum lighting that mimics natural variation in spectral composition from sunrise to sunset represents an important advance.
It is important to distinguish the spectral capacity of the NICU lighting system from lighting that varies color temperature over the course of a daytime period.Color temperature measures visual perception of light rather than actual spectral distribution.For example, the perception of violet can be created without the generation of wavelengths in the violet spectrum.Without violet photons, OPN5 stimulation will not occur.
At this early stage of translational research, there are many unanswered questions.The "dose" of light needed to activate OPN5 is not precisely defined.Preclinical studies demonstrate that sufficient photons reach deep brain structures in mice to stimulate OPN5 [11].Similarly, a growing body of studies documents the activation of OPN3 and OPN5 in a variety of non-ocular locations such as skin and adipose tissue [10,19].Opsins demonstrate absorption spectra over a range of wavelengths.For example, OPN5 can be stimulated at different efficiencies from 340-420 nm.The correlation of the absorption spectrum to biological effects in a real-world clinical setting will require additional study.
Our growing understanding of OPN3 and OPN5 implies their fundamental role in metabolic activities and energy homeostasis.Many conditions encountered in the NICU are associated with growth failure, such as gastroschisis, intestinal failure/short bowel, and bronchopulmonary dysplasia.The significance of body composition for neurodevelopmental outcomes is an area of increasing interest for the NICU population.Circadian influences on energy management are likely to be very relevant to common NICU practices such as enteral feeding protocols and parenteral nutrition.For example, time of day and exposure to relevant daylight wavelengths may profoundly influence how a neonate utilizes macronutrients and energy substrate.Post-discharge outcomes, such as visual acuity and neurobehavior, may also be impacted [17].
As we consider environmental light exposure in the context of neonatology practice, many questions require future study.The dose of daylight, including duration of exposure, intensity, and key wavelengths will require further refinement.This is particularly interesting when deep tissue structures expressing light-sensing opsins are considered.Electromagnetic wavelengths outside of the visual spectrum (infrared and ultraviolet) are also known to influence certain biological processes through non-opsin mediators.Our focus here is restricted to wavelengths within the visual spectrum known to activate human opsins.Finally, energy management, utilization of macronutrients, and growth are fundamental to survival.It is reasonable to expect that pathways regulated through opsin stimulation are relevant to human growth and development.Our NICU lighting system will now allow interrogation of these pathways, supporting the further optimization of growth and development during a crucial stage of life.Fig. 3 Weekly usage of the spectral lighting system was calculated from November 2021 to June 2023 for three randomly selected rooms.For each room, the percentage of spectral light usage was determined over the course of each week and recorded as a single data point.A progressive increase in use is shown over the measurement period.
intensity of photonic flux rises and falls to transitions from sunrise to midday to sunset.Key wavelengths corresponding to non-visual opsin activation vary throughout daylight progression.Violet and blue values dominate at sunrise and sunset.Artificial lighting conditions in the old and new NICUs (without the spectral lighting source) generated spectra of limited wavelength coverage consistent with known characteristics of artificial light sources.Notably, both old and new NICU (without the spectral lighting source) environments lacked wavelengths in the violet range capable of OPN5 activation (Fig.1).

Fig. 1
Fig. 1 Spectral distributions were measured with a ST-UV-100 microspectrometer (Ocean Insight, Orlando, FL) calibrated to a DH-3P-BAL-CAL light source.Panels A-D present representative spectral distribution measurements.Outdoor measurements (panels A and B) were taken with an integration time of 100 µs to prevent oversaturation.For indoor room measurements, the cosine corrector was placed at the level of the patient bed with an integration time of 100 ms.The spectral distribution of ambient lighting conditions (standard LED) in the University of Cincinnati Hospital ("Old") NICU is shown in panel C. Panel D depicts the spectral distribution of ambient lighting conditions including the spectral lighting system measured at the bedside in a NICU patient room at midday.The arrows in each panel designate the approximate peak activation wavelengths for OPN5, OPN3, and OPN4 (left to right).Violet and blue wavelengths dominate total photon delivery at dawn/dusk (A) compared to midday (B).Note the presence of violet photons in panels A, B, and D and the absence of violet photons in panel C.

Table 2 .
Implementation for clinical care.