- Email: [email protected]
In-Use Photostability Practice and Regulatory Evaluation for Pharmaceutical Products in an Age of Light-Emitting Diode Light Sources
Journal of Pharmaceutical Sciences xxx (2018) 1-5
Contents lists available at ScienceDirect
Journal of Pharmaceutical Sciences journal homepage: www.jpharmsci.org
Pharmaceutics, Drug Delivery and Pharmaceutical Technology
In-Use Photostability Practice and Regulatory Evaluation for Pharmaceutical Products in an Age of Light-Emitting Diode Light Sources Leonardo R. Allain*, Brittany C. Pierce, W. Peter Wuelfing, Allen C. Templeton, Roy Helmy MRL, Merck & Company, Inc., Kenilworth, New Jersey 07033
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 August 2018 Revised 27 September 2018 Accepted 3 October 2018
This article describes how the increased use of energy-efficient solid-state light sources (e.g., lightemitting diode [LED]-based illumination) in hospitals, pharmacies, and at home can help alleviate concerns of photodegradation for pharmaceuticals. LED light sources, unlike fluorescent ones, do not have spurious spectral contributions <400 nm. Because photostability is primarily evaluated in the International Council of Harmonization Q1B tests with older fluorescent bulb standards (International Organization for Standardization 10977), the amount of photodegradation observed can over-predict what happens in reality, as products are increasingly being stored and used in environments fitted with LED bulbs. Because photodegradation is premised on light absorption by a compound of interest (or a photosensitizer), one can use the overlap between the spectral distribution of a light source and the absorption spectra of a given compound to estimate if photodegradation is a possibility. Based on the absorption spectra of a sample of 150 pharmaceutical compounds in development, only 15% would meet the required overlap to be a candidate to undergo direct photodegradation in the presence of LED lights, against a baseline of 55% of compounds that would, when considering regular fluorescent lights. Biological drug products such as peptides and monoclonal antibodies are also expected to benefit from the use of more efficient solid-state lighting. © 2018 Published by Elsevier Inc. on behalf of the American Pharmacists Association.
Keywords: degradation product(s) photochemistry photodegradation regulatory science stability UV/Vis spectroscopy
Introduction The photostability risk for pharmaceutical products is primarily informed by International Council of Harmonization (ICH) Q1B testing protocols. These tests are designed to present conditions which are representative for a drug product under manufacturing, packaging, storage, and in-use lighting conditions. Although the primary package typically mitigates photostability risk in products during storage, their in-use stability, when not protected from light sources, is more challenging and is of note for parenteral formulations where drug is dissolved, allowing more rapid photochemical reactions. Also of note is that the majority of pharmaceutical compounds absorb light in the ultraviolet (UV) region, not the visible part of the spectra, limiting the opportunities for photoexcitation and reactivity to light absorption in the UV region. The way indoor environments are illuminated is
Declarations of interest: None. * Correspondence to: L.R. Allain (Telephone: þ1-215-652-7424). E-mail address: [email protected] (L.R. Allain).
rapidly changing with the mass adoption of light-emitting diode (LED) lighting versus fluorescent lighting for residential, pharmacy, and hospital use, and may modulate the photostability risk for pharmaceuticals. Fortunately, the spectral output of LED lights does not contain spurious contributions from UV wavelengths, and therefore lacks the ability to promote photoinstability of many drug substances or photosensitizers. Additionally, low emissivity windows, or those with glazing or reflective coating treatments, are being used as new architectural standards, further reducing UV exposure. We and other authors have previously carried out extensive data-driven studies to define the spectral output of light sources used in healthcare facilities to better define in-use environments common over the last decade.1-3 We describe here how the photostability risk is expected to decline, given these recent trends in lighting and architectural standards, and consideration of these changes should be given, when generating predictive stability results with current ICH Q1B testing protocols. Although created in the 1960s, LEDs only became economically practical for lighting purposes after the year 2000, with bulb installations reaching greater than 200 million in 2015.4 The U.S.
https://doi.org/10.1016/j.xphs.2018.10.003 0022-3549/© 2018 Published by Elsevier Inc. on behalf of the American Pharmacists Association.
2
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5
Department of Energy predicts that LEDs will account for 85% of all lighting installations by 2035.5 Diodes are solid-state devices that generate light by electron/hole recombination in semiconductor materials, distributed in a short bandwidth (20-50 nm), centered around a particular wavelength, and thus give the appearance of being monochromatic. Different doping schemes in the diodes allow the emission distribution to be extended toward higher wavelengths (lower energies), from 400 to 625 nm. LED lights are comprised of either a single diode with a phosphor to enable a broad spectral power distribution (SPD) or multiple diodes combined in 1 assembly without a phosphor (such as red, green, and blue).6 As such, blue-violet diodes yield the lowest wavelength output, and their emission cutoff is key to understand the amount of overlap that can occur with a drug substance’s absorption spectra, which would be able to promote any permissible photodegradation. In order to enhance energy efficiency, the blue diode light is centered on 430 nm, which leaves very little output below 400 nm, and no spurious UV emissions at lower wavelengths. This is in remarkable contrast to how fluorescent lights operate, where the primary source of light is the Hg emission lines in the UV, produced in the interior of the light bulb by excitation of mercury atoms in the gas phase. Most (but not all) of the UV-light energy thus produced is converted to visible wavelengths via phosphorescence of the materials present as a coating in the bulb, also known as the phosphors. Although emitted light is filtered through soda-lime glass, some UV radiation does escape (<400 nm). In fact some fluorescent lights are specifically designed to allow UV light through for sterilization or black light use (UV-C and UV-A, respectively). Given that the in-use period is indoors for the majority of non-topical pharmaceutical products, and particularly parenteral dosing, the changing of light source has a profound impact on stability. Companies involved in lighting solutions to
healthcare including Phillips and Cree promote LED lighting for multiple reasons, and there are some news reports of hospital groups making complete switch-overs to LED lighting in 2017.7-9 ICH photostability testing is carried out with 1 of the 2 options which are either a fluorescent source (International Organization for Standardization [ISO] 10977 compliant) and a separate UV source (typically a UV fluorescent lamp), or ID65 or D65 source emitting both UV and visible wavelengths in a combined UV and cool white visible lamp. Both options are used to predict photostability behavior associated to indoor light exposure, as well as window-filtered sunlight, whereas the ISO 10977-compliant or equivalent lamp (henceforth shortened to ISO 10977 in this article) is relevant to evaluate the indoor light photostability risk. However, the ISO 10977 has an SPD with several UV-A components, and may over-predict in-use clinical and home photodegradation, compared to LED lights. Low emissivity windows, with an effective UV cutoff, and the use of reflective coatings on windows are part of a growing trend in higher efficiency architectural standards, which may influence how we perceive photostability risk to pharmaceutical products during their out-of-package in-use period. A review of the transmissivity properties of different window panes show that spectrally selective UV-blocking and laminated glasses offer the most effective reduction in UV contributions, with UV cutoffs of 365 and 375 nm, respectively. Other modern panels, such as solar-reflective glass, and spectrally selective low emissivity, have cutoffs around 325 nm; uncoated, regular, window panes have a cutoff around 310 nm. The considerations for the choice of UV cutoff in a building project are not solely based on the cost of the glass panes, but also on the efficiency trade-off between heating costs in the winter (i.e., allowing more UV through, as well as infrared radiation) and cooling costs in the summer (blocking UV light, as well as infrared radiation).10
Figure 1. Overlap of the UV absorption spectra of compound A (yellow) with the spectral power distributions of an ISO 10977 fluorescent light (purple) and a white LED light (blue), and photodegradation of compound A.
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5
Methods and Results SPD measurements for Sylvania Octron and Cool White lamps were taken using a StellarNet Inc EPP 2000 spectroradiometer. A dark control background measurement was taken prior to SPD measurements of the various light sources. Figure 1 presents SPD measurements of a typical ISO 10977 fluorescent bulb versus a commercially available LED bulb. Photostability experiments were conducted by exposing model compounds to fluorescent and LED light at equal irradiances. Fluorescent light exposures were conducted in a Caron Model 6540-2 photostability chamber equipped with Sylvania cool white fluorescent lights (Sylvania Cool White, 18 W, F18T8/CW/K24 Hg) that are compliant with the ISO 10977 standard. LED light exposures were conducted in a home-built light box with a standard flush mount LED fixture installed. The active compound in Product A was considered to be photosensitive based on the overlap between the absorption spectrum of the compound and the emission spectrum of the fluorescent light source tested. Product A was exposed to 31,500 lux∙h based on a conservative assumption of an in-use stability of 2 days with room illumination at 1000 lux (16 h illumination per day). All experiments were conducted at ambient laboratory temperature and humidity. Dark control samples were wrapped in aluminum foil prior to exposure to fluorescent or LED lights. The assay results of photodegradation products are shown in Figure 1. Method for assay of degradation products in Product A: Samples were prepared in 90:10 methanol:water with subsequent dilution in 25:75 acetonitrile:water and analyzed by reverse-phase chromatography using UV detection at 300 nm on a Waters Acquity UPLC. Elution was performed on a Waters Acquity CSH column (100 2.1 mm, 1.7 mm) using a gradient of 0.1% phosphoric acid in water and acetonitrile. Discussion and Conclusions In Figure 1, the presence of emissions below 400 nm is striking in the ISO 10977 fluorescent bulb, with sharp peaks associated with electronic transitions of Hg or phosphors, and key UV contributions at approximately 410, 370, and 310 nm, which can drive photodegradation. The LED SPD is smooth, with a cutoff at 405 nm, with
3
no endogenous UV components below 400 nm. The pharmaceutical compound in Product A, whose absorption spectra is also depicted in Figure 1, undergoes extensive photodegradation under the ISO 10977 light source, with much smaller levels observed under exposure to LED white light (3.5% vs. 0.15% degradation, respectively). A key pre-condition to photodegradation is that absorption of light occurs for the specific electronic transition associated with photoreactivity.11 The integral overlap between UV absorption spectra is much more limited with the LED light, reducing the opportunities for populating an excited electronic state associated with the photodegradation process. As a general rule, the integral overlap between a compound’s absorption spectra and the light source power distribution offers a practical means to evaluate potential photodegradation susceptibility to a given light source. By applying this principle, a survey of 150 small-molecule pharmaceutical compounds from Merck & Company, Inc. (Kenilworth, NJ) reveals that only 15% have absorption spectra that would overlap with the LED spectral distribution, while 55% of these compounds have spectra that overlap with the fluorescent light spectral distribution. This coarse estimation does not take into account photodegradation via excipients as photosensitizers, but the operating principle is the same. For the minority of drug substances that are colored, thus having spectral absorption components in the visible range, it is possible that photodegradation behavior might be different depending on the type of white LED light used. For white LEDs that combine discrete diodes (blue, red, green), higher levels of degradation might occur if the colored drug substances were to have causative wavelengths that overlap with the output of one of the discrete diodes, compared with white LED lights that rely on a single diode and a phosphor. The ISO 10977 fluorescent bulb is an older standard (established in 1993 and revised in 2006 as ISO 18909), and more modern fluorescent bulbs are expected to have better phosphor composition and coverage, and thus be better engineered to more efficiently capture the UV Hg lines and transform them into visible radiation, improving their efficiency. This is shown in Figure 2, comparing the SPDs of the ISO 10977 standard against more modern fluorescent bulb counterparts, such as the Sylvania Octron, normalized to 1000 lux. However, modern fluorescent lights still have a considerable output below 400 nm, such as the compact fluorescent lamps shown in Figure 2.12
Figure 2. Spectral power distributions of commercially available fluorescent lights.
4
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5
Figure 3. The human eye photopic response curve. Comparisons between perceived illuminance from LED and fluorescent lights are based on the integral from 400 to 700 nm of the power distribution of the light source across this response curve.
Besides the absence of UV output, another key difference between LED sources and fluorescent ones is the total power of light emitted to achieve similar illuminance ratings, expressed in lux (lumen/m2). LEDs can achieve similar illuminance to fluorescent lights at a lower power emission, because their SPD more closely matches the human eye photopic response curve,13 shown in Figure 3. For example, commercial compact fluorescent lamps with an 800-lumen output are listed as 15 W, whereas white LED lights with the same luminosity output are listed as 9.5 W. As LED lights become more ubiquitous, considering their light emission properties and lower power distribution, the photostability risk posed by indoor lighting is expected to be reduced accordingly, and the photodegradation behavior observed using the ICH Q1B Option 2 might be over-represented, at least for developed countries that are rapidly adopting LED illumination. For photounstable products, the concept of a light budget is useful in framing how long a product should be studied and which illumination levels are representative to simulate its in-use conditions, for example, a lyophilized product that is reconstituted in a hospital environment, or for oral tablets that are allowed to be placed in calendar trays at home. In regards to representative illumination levels, the Illuminating Engineering Society published updated guidance on lightning for healthcare, such as in examination rooms, laboratories, procedure rooms, and pharmacies, where medications are likely to be used or stored.11 In these locations, the recommended lightning levels are 500-1000 lux. This is also in agreement with previous work done by Baertschi et al.1 in which direct measurements of illumination were performed in hospitals and pharmacies in the Philadelphia area. For example, taking into consideration the in-use testing conditions for oral products in a healthcare environment, a product allowed a 7-day in-use under indoor artificial lights would be exposed to 112 klux to simulate the worst case illuminance, using ICH Option 2 (ISO 10977).1-3 Any future changes in the regulatory landscape that acknowledges the shift to the use of LED lights may allow the potential to propose much longer in-use periods or more convenient pharmacy/hospital manipulation. Investigating the in-use photostability behavior with the 2 types of white LED lights (single diode with phosphor and multiple diodes) might be warranted, especially for compounds that have absorption contributions above 400 nm. Another progressing change in dwelling illumination is the use of reflective coatings on windows to improve energy efficiency in a building. UV reflective coatings on windows were first introduced over 40 years ago, and are now part of Leadership in Energy and Environmental Design building strategies, given their ability to
reduce UV transmittance by over 99% of total UV energy and significant infrared light rejection. The company 3M presents multiple case studies of hospital, pharmacy, and residential use for protective windows and window films at their web sites,14 showing the significant light rejection that these windows can afford, in some cases rejecting all light below 360 nm.15 Colored (tinted) glass also can further reject UV light. Installations of energy star windows, which are low emissivity, have doubled to over 80% from 2003 to 2010.16 This will also impact the extent of UV light that is available for absorption in-doors, and eliminate photodegradation for some products. A survey of windows to establish potentially a new light standard may be necessary as time progresses. From a regulatory perspective, in-use stability studies that evaluate exposure to indirect sunlight (ICH Q1B Option 1 or the UV portion of Option 2) are likely to be over-predictive, considering environments where modern window panes with a larger UV rejection are installed. However, the base-case scenario involving regular glass windows is still applicable and relevant, especially for geographic areas with low rates of adoption of new building standards. Although this article has focused on small molecules, the same principles are also important for biologic compounds, given that many common amino acids (tryptophan, tyrosine, and histidine) absorb wavelengths beyond 320 nm17,18; monoclonal antibodies can also be detrimentally impacted by photostability upon UV exposure, with breakage of di-sulfide bonds and changes in the tertiary structure of monoclonal antibodies or aggregation.19-21 LEDs are thus a much more favorable option for controlling photostability of pharmaceutical products, and their current accelerated adoption will allow longer in-use periods for products of particular concern. Solid-state LED lights are expected to become the future standard for artificial illumination in a variety of settings, including hospitals, operating rooms, pharmacies, and at home. In our estimates, at least 85% of small-molecule pharmaceutical compounds would not photodegrade under LED lights. Acknowledgments The authors would like to thank Dr Lee Klein and Dr Olivier Mozziconacci for their thoughtful discussion, and Merck Research Laboratories for supporting this work. References 1. Baertschi SW, Clapham D, Foti C, et al. Implications of in-use photostability: proposed guidance for photostability testing and labeling to support the administration of photosensitive pharmaceutical products, part 1: drug products administered by injection. J Pharm Sci. 2013;102(11):3888-3899. 2. Baertschi SW, Clapham D, Foti C, et al. Implications of in-use photostability: proposed guidance for photostability testing and labeling to support the administration of photosensitive pharmaceutical products, part 2: topical drug products. J Pharm Sci. 2015;104(9):2688-2701. 3. Allain L, Baertschi SW, Clapham D, et al. Implications of in-use photostability: proposed guidance for photostability testing and labeling to support the administration of photosensitive pharmaceutical products, part 3: oral drug products. J Pharm Sci. 2016;105(5):1586-1594. 4. Laskow S. Inventing the LED Lightbulb. The Atlantic. Available at: www. theatlantic.com/technology/archive/2014/09/who-invented-the-newlightbulb/ 379905/. Accessed September 10, 2014. 5. Donohoo-Vallett P, Gilman P, Feldman D, et al. The future arrives for five clean energy technologies. U.S. Department of Energy: Washington, DC; 2016. Available at: https://energy.gov/sites/prod/files/2016/09/f33/Revolutiona% CC%82%E2%82%ACNow%202016%20Report_2.pdf. Accessed August 1, 2018. 6. Nardelli A, Deuschle E, Azevedo LD, Pessoa JLN, Ghisi E. Assessment of Light Emitting Diodes technology for general lighting: a critical review. Renew Sustain Energy Rev. 2017;75:368-379. 7. Designing people-centric hospitals using Philips lighting solutions. Document order number: HEALTHCAREAPPGUIDE#2e5/14 INT. Available at: http://www. mea.lighting.philips.com/b-dam/b2b-li/en_AA/Applications/masthead-pdfs/ Healthcare-Application-Guide.pdf. Accessed August 1, 2018. 8. Health care facilities lightening case studies. Cree: Durham, NC; 2014. Available at: http://lighting.cree.com/applications/healthcare. Accessed August 1, 2018.
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5 9. McLaren health care selects Cree LED lighting to deliver intelligent light across its statewide hospital network. Available at: http://www.cree.com/news-media/ news/article/mclaren-health-care-selects-cree-led-lighting-to-deliver-intelligentlight-across-its-statewide-hospital-network. Accessed February 21, 2017. 10. DiLaura D, Houser H, Mistrick R, Steffy G. Lighting Handbook. 10th ed. New York, NY: Illuminating Engineering Society; 2011. 11. Ahmad I, Ahmed S, Anwar Z, Sheraz MA, Sikorski M. Photostability and photostabilization of drugs and drug products. Int J Photoenergy. 2016;2016:19. Article ID 8135608. 12. SPDs from CFL and FL lights presented are available from the SPD Database (LSPDD). Available at: http://galileo.graphycs.cegepsherbrooke.qc.ca/app/en/lamps. Accessed August 1, 2018. 13. Thapan K, Arendt J, Skene DJ. An action spectrum for melatonin suppression: evidence for a novel non-rod, non-cone photoreceptor system in humans. J Physiol. 2001;535:261-267. 14. 3M Thinsulate Window Film, smarter windows brighter you. Maplewood, MN: 3M; 2015. Available at: http://multimedia.3m.com/mws/media/1142269O/3mthinsulate-window-film-commercial-brochure.pdf?fn¼98-0150-0662-4_Single %20page.pdf. Accessed August 1, 2018. 15. Arasteh D, Carmody J, Heschong L, Selkowitz S. Residential Windows: A Guide to New Technologies and Energy Performance. New York, NY: W. W. Norton & Company; 2007:91.
5
16. Rissman J, Kennan H. Case studies on the government’s role in energy technology innovationdlow-emissivity windows. Washington, DC: American Energy Innovation Council; 2013. Available at: http://americanenergyinnovation.org/wp-content/ uploads/2013/03/Case-Low-e-Windows.pdf. Accessed August 1, 2018. 17. Creed D. The photophysics and photochemistry of the near-UV absorbing amino-acids. 1. Tryptophan and its simple derivatives. Photochem Photobiol. 1984;39(4):537-562. n LB, Scho € neich C. Solid-state pho18. Miller BL, Hageman MJ, Thamann TJ, Barro todegradation of bovine somatotropin (bovine growth hormone): evidence for tryptophan-mediated photooxidation of disulfide bonds. J Pharm Sci. 2003;92(8):1698-1709. €neich C. Effect of conformation on the photodegradation 19. Mozziconacci O, Scho of Trp- and cystine-containing cyclic peptides: octreotide somatostatin. Mol Pharm. 2014;11(10):3537-3546. €neich C. Light20. Haywood J, Mozziconacci O, Allegre KM, Kerwin BA, Scho induced conversion of Trp to Gly and Gly hydroperoxide in IgG1. Mol Pharm. 2013;10(3):1146-1150. € neich C. The photolysis of di21. Zhou S, Mozziconacci O, Kerwin BA, Sch o sulfide bonds in IgG1 and IgG2 leads to selective intramolecular hydrogen transfer reactions of cysteine thiyl radicals, probed by covalent h/d exchange and RPLC-MS/MS analysis. Pharm Res. 2013;30(5):12911299.
Contents lists available at ScienceDirect
Journal of Pharmaceutical Sciences journal homepage: www.jpharmsci.org
Pharmaceutics, Drug Delivery and Pharmaceutical Technology
In-Use Photostability Practice and Regulatory Evaluation for Pharmaceutical Products in an Age of Light-Emitting Diode Light Sources Leonardo R. Allain*, Brittany C. Pierce, W. Peter Wuelfing, Allen C. Templeton, Roy Helmy MRL, Merck & Company, Inc., Kenilworth, New Jersey 07033
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 August 2018 Revised 27 September 2018 Accepted 3 October 2018
This article describes how the increased use of energy-efficient solid-state light sources (e.g., lightemitting diode [LED]-based illumination) in hospitals, pharmacies, and at home can help alleviate concerns of photodegradation for pharmaceuticals. LED light sources, unlike fluorescent ones, do not have spurious spectral contributions <400 nm. Because photostability is primarily evaluated in the International Council of Harmonization Q1B tests with older fluorescent bulb standards (International Organization for Standardization 10977), the amount of photodegradation observed can over-predict what happens in reality, as products are increasingly being stored and used in environments fitted with LED bulbs. Because photodegradation is premised on light absorption by a compound of interest (or a photosensitizer), one can use the overlap between the spectral distribution of a light source and the absorption spectra of a given compound to estimate if photodegradation is a possibility. Based on the absorption spectra of a sample of 150 pharmaceutical compounds in development, only 15% would meet the required overlap to be a candidate to undergo direct photodegradation in the presence of LED lights, against a baseline of 55% of compounds that would, when considering regular fluorescent lights. Biological drug products such as peptides and monoclonal antibodies are also expected to benefit from the use of more efficient solid-state lighting. © 2018 Published by Elsevier Inc. on behalf of the American Pharmacists Association.
Keywords: degradation product(s) photochemistry photodegradation regulatory science stability UV/Vis spectroscopy
Introduction The photostability risk for pharmaceutical products is primarily informed by International Council of Harmonization (ICH) Q1B testing protocols. These tests are designed to present conditions which are representative for a drug product under manufacturing, packaging, storage, and in-use lighting conditions. Although the primary package typically mitigates photostability risk in products during storage, their in-use stability, when not protected from light sources, is more challenging and is of note for parenteral formulations where drug is dissolved, allowing more rapid photochemical reactions. Also of note is that the majority of pharmaceutical compounds absorb light in the ultraviolet (UV) region, not the visible part of the spectra, limiting the opportunities for photoexcitation and reactivity to light absorption in the UV region. The way indoor environments are illuminated is
Declarations of interest: None. * Correspondence to: L.R. Allain (Telephone: þ1-215-652-7424). E-mail address: [email protected] (L.R. Allain).
rapidly changing with the mass adoption of light-emitting diode (LED) lighting versus fluorescent lighting for residential, pharmacy, and hospital use, and may modulate the photostability risk for pharmaceuticals. Fortunately, the spectral output of LED lights does not contain spurious contributions from UV wavelengths, and therefore lacks the ability to promote photoinstability of many drug substances or photosensitizers. Additionally, low emissivity windows, or those with glazing or reflective coating treatments, are being used as new architectural standards, further reducing UV exposure. We and other authors have previously carried out extensive data-driven studies to define the spectral output of light sources used in healthcare facilities to better define in-use environments common over the last decade.1-3 We describe here how the photostability risk is expected to decline, given these recent trends in lighting and architectural standards, and consideration of these changes should be given, when generating predictive stability results with current ICH Q1B testing protocols. Although created in the 1960s, LEDs only became economically practical for lighting purposes after the year 2000, with bulb installations reaching greater than 200 million in 2015.4 The U.S.
https://doi.org/10.1016/j.xphs.2018.10.003 0022-3549/© 2018 Published by Elsevier Inc. on behalf of the American Pharmacists Association.
2
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5
Department of Energy predicts that LEDs will account for 85% of all lighting installations by 2035.5 Diodes are solid-state devices that generate light by electron/hole recombination in semiconductor materials, distributed in a short bandwidth (20-50 nm), centered around a particular wavelength, and thus give the appearance of being monochromatic. Different doping schemes in the diodes allow the emission distribution to be extended toward higher wavelengths (lower energies), from 400 to 625 nm. LED lights are comprised of either a single diode with a phosphor to enable a broad spectral power distribution (SPD) or multiple diodes combined in 1 assembly without a phosphor (such as red, green, and blue).6 As such, blue-violet diodes yield the lowest wavelength output, and their emission cutoff is key to understand the amount of overlap that can occur with a drug substance’s absorption spectra, which would be able to promote any permissible photodegradation. In order to enhance energy efficiency, the blue diode light is centered on 430 nm, which leaves very little output below 400 nm, and no spurious UV emissions at lower wavelengths. This is in remarkable contrast to how fluorescent lights operate, where the primary source of light is the Hg emission lines in the UV, produced in the interior of the light bulb by excitation of mercury atoms in the gas phase. Most (but not all) of the UV-light energy thus produced is converted to visible wavelengths via phosphorescence of the materials present as a coating in the bulb, also known as the phosphors. Although emitted light is filtered through soda-lime glass, some UV radiation does escape (<400 nm). In fact some fluorescent lights are specifically designed to allow UV light through for sterilization or black light use (UV-C and UV-A, respectively). Given that the in-use period is indoors for the majority of non-topical pharmaceutical products, and particularly parenteral dosing, the changing of light source has a profound impact on stability. Companies involved in lighting solutions to
healthcare including Phillips and Cree promote LED lighting for multiple reasons, and there are some news reports of hospital groups making complete switch-overs to LED lighting in 2017.7-9 ICH photostability testing is carried out with 1 of the 2 options which are either a fluorescent source (International Organization for Standardization [ISO] 10977 compliant) and a separate UV source (typically a UV fluorescent lamp), or ID65 or D65 source emitting both UV and visible wavelengths in a combined UV and cool white visible lamp. Both options are used to predict photostability behavior associated to indoor light exposure, as well as window-filtered sunlight, whereas the ISO 10977-compliant or equivalent lamp (henceforth shortened to ISO 10977 in this article) is relevant to evaluate the indoor light photostability risk. However, the ISO 10977 has an SPD with several UV-A components, and may over-predict in-use clinical and home photodegradation, compared to LED lights. Low emissivity windows, with an effective UV cutoff, and the use of reflective coatings on windows are part of a growing trend in higher efficiency architectural standards, which may influence how we perceive photostability risk to pharmaceutical products during their out-of-package in-use period. A review of the transmissivity properties of different window panes show that spectrally selective UV-blocking and laminated glasses offer the most effective reduction in UV contributions, with UV cutoffs of 365 and 375 nm, respectively. Other modern panels, such as solar-reflective glass, and spectrally selective low emissivity, have cutoffs around 325 nm; uncoated, regular, window panes have a cutoff around 310 nm. The considerations for the choice of UV cutoff in a building project are not solely based on the cost of the glass panes, but also on the efficiency trade-off between heating costs in the winter (i.e., allowing more UV through, as well as infrared radiation) and cooling costs in the summer (blocking UV light, as well as infrared radiation).10
Figure 1. Overlap of the UV absorption spectra of compound A (yellow) with the spectral power distributions of an ISO 10977 fluorescent light (purple) and a white LED light (blue), and photodegradation of compound A.
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5
Methods and Results SPD measurements for Sylvania Octron and Cool White lamps were taken using a StellarNet Inc EPP 2000 spectroradiometer. A dark control background measurement was taken prior to SPD measurements of the various light sources. Figure 1 presents SPD measurements of a typical ISO 10977 fluorescent bulb versus a commercially available LED bulb. Photostability experiments were conducted by exposing model compounds to fluorescent and LED light at equal irradiances. Fluorescent light exposures were conducted in a Caron Model 6540-2 photostability chamber equipped with Sylvania cool white fluorescent lights (Sylvania Cool White, 18 W, F18T8/CW/K24 Hg) that are compliant with the ISO 10977 standard. LED light exposures were conducted in a home-built light box with a standard flush mount LED fixture installed. The active compound in Product A was considered to be photosensitive based on the overlap between the absorption spectrum of the compound and the emission spectrum of the fluorescent light source tested. Product A was exposed to 31,500 lux∙h based on a conservative assumption of an in-use stability of 2 days with room illumination at 1000 lux (16 h illumination per day). All experiments were conducted at ambient laboratory temperature and humidity. Dark control samples were wrapped in aluminum foil prior to exposure to fluorescent or LED lights. The assay results of photodegradation products are shown in Figure 1. Method for assay of degradation products in Product A: Samples were prepared in 90:10 methanol:water with subsequent dilution in 25:75 acetonitrile:water and analyzed by reverse-phase chromatography using UV detection at 300 nm on a Waters Acquity UPLC. Elution was performed on a Waters Acquity CSH column (100 2.1 mm, 1.7 mm) using a gradient of 0.1% phosphoric acid in water and acetonitrile. Discussion and Conclusions In Figure 1, the presence of emissions below 400 nm is striking in the ISO 10977 fluorescent bulb, with sharp peaks associated with electronic transitions of Hg or phosphors, and key UV contributions at approximately 410, 370, and 310 nm, which can drive photodegradation. The LED SPD is smooth, with a cutoff at 405 nm, with
3
no endogenous UV components below 400 nm. The pharmaceutical compound in Product A, whose absorption spectra is also depicted in Figure 1, undergoes extensive photodegradation under the ISO 10977 light source, with much smaller levels observed under exposure to LED white light (3.5% vs. 0.15% degradation, respectively). A key pre-condition to photodegradation is that absorption of light occurs for the specific electronic transition associated with photoreactivity.11 The integral overlap between UV absorption spectra is much more limited with the LED light, reducing the opportunities for populating an excited electronic state associated with the photodegradation process. As a general rule, the integral overlap between a compound’s absorption spectra and the light source power distribution offers a practical means to evaluate potential photodegradation susceptibility to a given light source. By applying this principle, a survey of 150 small-molecule pharmaceutical compounds from Merck & Company, Inc. (Kenilworth, NJ) reveals that only 15% have absorption spectra that would overlap with the LED spectral distribution, while 55% of these compounds have spectra that overlap with the fluorescent light spectral distribution. This coarse estimation does not take into account photodegradation via excipients as photosensitizers, but the operating principle is the same. For the minority of drug substances that are colored, thus having spectral absorption components in the visible range, it is possible that photodegradation behavior might be different depending on the type of white LED light used. For white LEDs that combine discrete diodes (blue, red, green), higher levels of degradation might occur if the colored drug substances were to have causative wavelengths that overlap with the output of one of the discrete diodes, compared with white LED lights that rely on a single diode and a phosphor. The ISO 10977 fluorescent bulb is an older standard (established in 1993 and revised in 2006 as ISO 18909), and more modern fluorescent bulbs are expected to have better phosphor composition and coverage, and thus be better engineered to more efficiently capture the UV Hg lines and transform them into visible radiation, improving their efficiency. This is shown in Figure 2, comparing the SPDs of the ISO 10977 standard against more modern fluorescent bulb counterparts, such as the Sylvania Octron, normalized to 1000 lux. However, modern fluorescent lights still have a considerable output below 400 nm, such as the compact fluorescent lamps shown in Figure 2.12
Figure 2. Spectral power distributions of commercially available fluorescent lights.
4
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5
Figure 3. The human eye photopic response curve. Comparisons between perceived illuminance from LED and fluorescent lights are based on the integral from 400 to 700 nm of the power distribution of the light source across this response curve.
Besides the absence of UV output, another key difference between LED sources and fluorescent ones is the total power of light emitted to achieve similar illuminance ratings, expressed in lux (lumen/m2). LEDs can achieve similar illuminance to fluorescent lights at a lower power emission, because their SPD more closely matches the human eye photopic response curve,13 shown in Figure 3. For example, commercial compact fluorescent lamps with an 800-lumen output are listed as 15 W, whereas white LED lights with the same luminosity output are listed as 9.5 W. As LED lights become more ubiquitous, considering their light emission properties and lower power distribution, the photostability risk posed by indoor lighting is expected to be reduced accordingly, and the photodegradation behavior observed using the ICH Q1B Option 2 might be over-represented, at least for developed countries that are rapidly adopting LED illumination. For photounstable products, the concept of a light budget is useful in framing how long a product should be studied and which illumination levels are representative to simulate its in-use conditions, for example, a lyophilized product that is reconstituted in a hospital environment, or for oral tablets that are allowed to be placed in calendar trays at home. In regards to representative illumination levels, the Illuminating Engineering Society published updated guidance on lightning for healthcare, such as in examination rooms, laboratories, procedure rooms, and pharmacies, where medications are likely to be used or stored.11 In these locations, the recommended lightning levels are 500-1000 lux. This is also in agreement with previous work done by Baertschi et al.1 in which direct measurements of illumination were performed in hospitals and pharmacies in the Philadelphia area. For example, taking into consideration the in-use testing conditions for oral products in a healthcare environment, a product allowed a 7-day in-use under indoor artificial lights would be exposed to 112 klux to simulate the worst case illuminance, using ICH Option 2 (ISO 10977).1-3 Any future changes in the regulatory landscape that acknowledges the shift to the use of LED lights may allow the potential to propose much longer in-use periods or more convenient pharmacy/hospital manipulation. Investigating the in-use photostability behavior with the 2 types of white LED lights (single diode with phosphor and multiple diodes) might be warranted, especially for compounds that have absorption contributions above 400 nm. Another progressing change in dwelling illumination is the use of reflective coatings on windows to improve energy efficiency in a building. UV reflective coatings on windows were first introduced over 40 years ago, and are now part of Leadership in Energy and Environmental Design building strategies, given their ability to
reduce UV transmittance by over 99% of total UV energy and significant infrared light rejection. The company 3M presents multiple case studies of hospital, pharmacy, and residential use for protective windows and window films at their web sites,14 showing the significant light rejection that these windows can afford, in some cases rejecting all light below 360 nm.15 Colored (tinted) glass also can further reject UV light. Installations of energy star windows, which are low emissivity, have doubled to over 80% from 2003 to 2010.16 This will also impact the extent of UV light that is available for absorption in-doors, and eliminate photodegradation for some products. A survey of windows to establish potentially a new light standard may be necessary as time progresses. From a regulatory perspective, in-use stability studies that evaluate exposure to indirect sunlight (ICH Q1B Option 1 or the UV portion of Option 2) are likely to be over-predictive, considering environments where modern window panes with a larger UV rejection are installed. However, the base-case scenario involving regular glass windows is still applicable and relevant, especially for geographic areas with low rates of adoption of new building standards. Although this article has focused on small molecules, the same principles are also important for biologic compounds, given that many common amino acids (tryptophan, tyrosine, and histidine) absorb wavelengths beyond 320 nm17,18; monoclonal antibodies can also be detrimentally impacted by photostability upon UV exposure, with breakage of di-sulfide bonds and changes in the tertiary structure of monoclonal antibodies or aggregation.19-21 LEDs are thus a much more favorable option for controlling photostability of pharmaceutical products, and their current accelerated adoption will allow longer in-use periods for products of particular concern. Solid-state LED lights are expected to become the future standard for artificial illumination in a variety of settings, including hospitals, operating rooms, pharmacies, and at home. In our estimates, at least 85% of small-molecule pharmaceutical compounds would not photodegrade under LED lights. Acknowledgments The authors would like to thank Dr Lee Klein and Dr Olivier Mozziconacci for their thoughtful discussion, and Merck Research Laboratories for supporting this work. References 1. Baertschi SW, Clapham D, Foti C, et al. Implications of in-use photostability: proposed guidance for photostability testing and labeling to support the administration of photosensitive pharmaceutical products, part 1: drug products administered by injection. J Pharm Sci. 2013;102(11):3888-3899. 2. Baertschi SW, Clapham D, Foti C, et al. Implications of in-use photostability: proposed guidance for photostability testing and labeling to support the administration of photosensitive pharmaceutical products, part 2: topical drug products. J Pharm Sci. 2015;104(9):2688-2701. 3. Allain L, Baertschi SW, Clapham D, et al. Implications of in-use photostability: proposed guidance for photostability testing and labeling to support the administration of photosensitive pharmaceutical products, part 3: oral drug products. J Pharm Sci. 2016;105(5):1586-1594. 4. Laskow S. Inventing the LED Lightbulb. The Atlantic. Available at: www. theatlantic.com/technology/archive/2014/09/who-invented-the-newlightbulb/ 379905/. Accessed September 10, 2014. 5. Donohoo-Vallett P, Gilman P, Feldman D, et al. The future arrives for five clean energy technologies. U.S. Department of Energy: Washington, DC; 2016. Available at: https://energy.gov/sites/prod/files/2016/09/f33/Revolutiona% CC%82%E2%82%ACNow%202016%20Report_2.pdf. Accessed August 1, 2018. 6. Nardelli A, Deuschle E, Azevedo LD, Pessoa JLN, Ghisi E. Assessment of Light Emitting Diodes technology for general lighting: a critical review. Renew Sustain Energy Rev. 2017;75:368-379. 7. Designing people-centric hospitals using Philips lighting solutions. Document order number: HEALTHCAREAPPGUIDE#2e5/14 INT. Available at: http://www. mea.lighting.philips.com/b-dam/b2b-li/en_AA/Applications/masthead-pdfs/ Healthcare-Application-Guide.pdf. Accessed August 1, 2018. 8. Health care facilities lightening case studies. Cree: Durham, NC; 2014. Available at: http://lighting.cree.com/applications/healthcare. Accessed August 1, 2018.
L.R. Allain et al. / Journal of Pharmaceutical Sciences xxx (2018) 1-5 9. McLaren health care selects Cree LED lighting to deliver intelligent light across its statewide hospital network. Available at: http://www.cree.com/news-media/ news/article/mclaren-health-care-selects-cree-led-lighting-to-deliver-intelligentlight-across-its-statewide-hospital-network. Accessed February 21, 2017. 10. DiLaura D, Houser H, Mistrick R, Steffy G. Lighting Handbook. 10th ed. New York, NY: Illuminating Engineering Society; 2011. 11. Ahmad I, Ahmed S, Anwar Z, Sheraz MA, Sikorski M. Photostability and photostabilization of drugs and drug products. Int J Photoenergy. 2016;2016:19. Article ID 8135608. 12. SPDs from CFL and FL lights presented are available from the SPD Database (LSPDD). Available at: http://galileo.graphycs.cegepsherbrooke.qc.ca/app/en/lamps. Accessed August 1, 2018. 13. Thapan K, Arendt J, Skene DJ. An action spectrum for melatonin suppression: evidence for a novel non-rod, non-cone photoreceptor system in humans. J Physiol. 2001;535:261-267. 14. 3M Thinsulate Window Film, smarter windows brighter you. Maplewood, MN: 3M; 2015. Available at: http://multimedia.3m.com/mws/media/1142269O/3mthinsulate-window-film-commercial-brochure.pdf?fn¼98-0150-0662-4_Single %20page.pdf. Accessed August 1, 2018. 15. Arasteh D, Carmody J, Heschong L, Selkowitz S. Residential Windows: A Guide to New Technologies and Energy Performance. New York, NY: W. W. Norton & Company; 2007:91.
5
16. Rissman J, Kennan H. Case studies on the government’s role in energy technology innovationdlow-emissivity windows. Washington, DC: American Energy Innovation Council; 2013. Available at: http://americanenergyinnovation.org/wp-content/ uploads/2013/03/Case-Low-e-Windows.pdf. Accessed August 1, 2018. 17. Creed D. The photophysics and photochemistry of the near-UV absorbing amino-acids. 1. Tryptophan and its simple derivatives. Photochem Photobiol. 1984;39(4):537-562. n LB, Scho € neich C. Solid-state pho18. Miller BL, Hageman MJ, Thamann TJ, Barro todegradation of bovine somatotropin (bovine growth hormone): evidence for tryptophan-mediated photooxidation of disulfide bonds. J Pharm Sci. 2003;92(8):1698-1709. €neich C. Effect of conformation on the photodegradation 19. Mozziconacci O, Scho of Trp- and cystine-containing cyclic peptides: octreotide somatostatin. Mol Pharm. 2014;11(10):3537-3546. €neich C. Light20. Haywood J, Mozziconacci O, Allegre KM, Kerwin BA, Scho induced conversion of Trp to Gly and Gly hydroperoxide in IgG1. Mol Pharm. 2013;10(3):1146-1150. € neich C. The photolysis of di21. Zhou S, Mozziconacci O, Kerwin BA, Sch o sulfide bonds in IgG1 and IgG2 leads to selective intramolecular hydrogen transfer reactions of cysteine thiyl radicals, probed by covalent h/d exchange and RPLC-MS/MS analysis. Pharm Res. 2013;30(5):12911299.