Photodynamic therapy improves skin antisepsis as a prevention strategy in arthroplasty procedures: A pilot study

: Background: Current standard skin antisepsis to prevent surgical site infections are ineffective to eradicate all skin-colonizing bacteria. Photodynamic therapy (PDT) has shown bactericidal effects in vitro, but no clinical study with improvements in skin antisepsis has been documented. Methods: We investigated the effect of methyl aminolevulinate (MAL)-PDT versus no PDT for skin antisepsis treatment (povidone-iodine/alcohol) in the groin of 10 healthy participants. Skin swabs were taken at baseline, immediately after PDT, and after skin antisepsis treatment to cultivate bacteria. At day 7 and 21, bacterial cultures were repeated before and after antisepsis treatment without PDT. Skin biopsies were performed to examine the grade of inflammation. Results: Skin-colonizing bacteria were found in all 20 participants at baseline sampling. Immediately after MAL-PDT, skin was sterile in 7 (70%) participants before and in all 10 (100%) participants after skin antisepsis treatment. In contrast, we found skin-colonizing bacteria in 5 (50%) participants of the control group receiving only skin antisepsis. After 7 and 21 days, skin sterility was similar to the baseline. We observed slight perivascular inflammation with lymphocytes and eosinophils without changes in the histomorphology of eccrine or sebaceous glands in skin biopsies. PDT was generally well tolerated except for localized redness. Conclusion: MAL-PDT with skin antisepsis treatment sterilized skin immediately after its use but did not maintain sterility 7-21 days post-treatment. Due to local side effects, further clinical studies with less intensive PDT conditions or other photosensitizers are needed before PDT is integrated into Background: Current standard skin antisepsis to prevent surgical site infections are ine ﬀ ective to eradicate all skin-colonizing bacteria. Photodynamic therapy (PDT) has shown bactericidal e ﬀ ects in vitro, but no clinical study with improvements in skin antisepsis has been documented. Methods: We investigated the e ﬀ ect of methyl aminolevulinate (MAL)-PDT versus no PDT for skin antisepsis treatment (povidone-iodine/alcohol) in the groin of 10 healthy participants. Skin swabs were taken at baseline, immediately after PDT, and after skin antisepsis treatment to cultivate bacteria. At day 7 and 21, bacterial cultures were repeated before and after antisepsis treatment without PDT. Skin biopsies were performed to examine the grade of in ﬂ ammation. Results: Skin-colonizing bacteria were found in all 20 participants at baseline sampling. Immediately after MAL- PDT, skin was sterile in 7 (70%) participants before and in all 10 (100%) participants after skin antisepsis treatment. In contrast, we found skin-colonizing bacteria in 5 (50%) participants of the control group receiving only skin antisepsis. After 7 and 21 days, skin sterility was similar to the baseline. We observed slight perivascular in ﬂ ammation with lymphocytes and eosinophils without changes in the histomorphology of eccrine or sebaceous glands in skin biopsies. PDT was generally well tolerated except for localized redness. Conclusion: MAL-PDT with skin antisepsis treatment sterilized skin immediately after its use but did not maintain sterility 7 – 21 days post-treatment. Due to local side e ﬀ ects, further clinical studies with less intensive PDT conditions or other photosensitizers are needed before PDT is integrated into clinical practice.


Introduction
The use of orthopedic implants has been steadily rising over the last decade due to the growing demand for periprosthetic joints in the expanding elderly population. Periprosthetic joint infections (PJI) are one of the most feared complications after surgical intervention as these infections have biofilm etiologies that are difficult to treat and impact the quality of life, increasing health-care costs [1]. Within a mature biofilm, bacteria are protected against most current antibiotics by a coating of exopolysaccharide matrix, preventing antibiotic penetration and by bacterial heterogeneity, decreasing available antibiotic targets within bacteria. The most commonly isolated microorganisms in PJI are staphylococci, streptococci, enterococci, Gram-negative bacteria, and anaerobic bacteria such as Cutibacterium acnes/ Cutibacterium avidum (formerly Propionibacterium acnes/ Propionibacterium avidum) [ 1,2], all of which may inhabit the normal skin flora.
A majority of PJI are acquired during surgery, with the same bacteria that colonize the skin surface serving as infectious invaders of the orthopedic implant [3]. Skin antisepsis at incision sites should be the primary intervention to prevent commensal bacterial infection into deep tissues during surgery. Recent studies have shown that the current standards for skin antisepsis are not effective enough to sterilize the skin before arthroplasty [4]. Lee et al. showed a significant growth rate of C. acnes in 7 out of 10 volunteers with skin biopsies, despite skin preparation with chlorhexidine/alcohol prior to sampling [5]. Maurer et al. detected bacterial growth of skin commensals after standard https://doi.org/10.1016/j.pdpdt.2020.101941 Received 6 July 2020; Received in revised form 16 July 2020; Accepted 27 July 2020 antisepsis in the groin in 50% of the patients (manuscript in review, oral communication FP 81 at the European Bone and Joint Infection Society (EBJIS) meeting, Antwerp, 2019). Different approaches to improve antisepsis have been examined throughout the years, some with no effects [4] and others with promising results [6,7]. Local antibiotics risk the emergence of bacterial resistance, making this approach undesirable in daily surgical practice [8].
The antimicrobial effect of photodynamic therapy (PDT) has already been established in previous studies and relies on photophysical and photochemical reactions to eliminate bacteria [9,10]. We hypothesize that PDT reduces the growth of skin-colonizing bacteria when combined with skin antisepsis treatment, the latter already being routinely performed prior to arthroplasty surgeries. Therefore, we evaluated the effect of PDT with methyl aminolevulinate (MAL) as the prodrug for the photosensitizer protoporphyrin IX (Pp IX) on skin-colonizing bacteria in a pilot study with 10 healthy participants. Promising results will provide the basis for planning a future clinical study in an orthopedic center. Since PDT triggers subtle skin inflammation which may be suboptimal for immediate surgery, we evaluated the effect of PDT at baseline, immediately after PDT and skin antisepsis treatment, and after 7 and 21 days analyzing bacterial growth and severity of skin inflammation.

Participants and study design
In this single center clinical trial, healthy male and female participants (≥ 18 years) were recruited to have PDT in combination with skin antisepsis. The study was approved by the local ethical committee (BASEC Nr 2019-00252) and an informed consent was obtained from each participant prior to treatment. From September-November 2019, we enrolled 10 healthy participants for PDT and another 10 healthy participants for the control group to be treated with skin antisepsis only. We excluded pregnant and lactating women, participants unable to follow instructions of the study leaders (e.g. language problems, psychological disorders, dementia, etc), taking antibiotics 14 days prior to PDT or until follow-up at day 14, receiving oral retinoid therapy within the last 6 months or anti-inflammatory agents such as Non-Steroidal Anti-Inflammatory Drugs within 14 days prior and after PDT, taking any photosensitizing drugs within 4 weeks prior to PDT, and those who had a history of photosensitive disorder or Fitzpatrick's skin phototype VeVI.
Participant characteristics such as age, BMI, comorbidities, allergies, and medication were recorded for all 20 participants on the first day of their clinical visit. The study design for both groups is shown in Fig. 1. For all 10 participants receiving PDT treatment, skin swabs were immediately taken before and after PDT and after standard skin antisepsis treatment on days 0, 7, and 21 to investigate the amount of bacterial species on human skin. All study participants had follow-up examinations to evaluate the extent of skin inflammation after PDT. Participants were specifically asked about any redness, pain, skin irritation and itching, other abnormalities, and the need for medications for relief from symptoms. In a subgroup of patients, additional dermis biopsies of the treated groin area were taken immediately after PDT, after 7 days, and again after 21 days. A biopsy from the opposite, untreated side of the body was used as a control.

Skin swabs
Skin swabs were taken from the right groin area, immediately below the anterior superior iliac spine with sterile blades scraping at a 45°a ngle. Material was removed carefully from the blade and inoculated into an eSwab Collection and Transport System (Copan, Italy). All swabs were analyzed for bacterial growth at the Institute of Medical Microbiology of the University of Zurich. All skin swabs (1-7) were carried out using the same technique.

Photodynamic therapy (PDT)
PDT requires three elements: a photosensitizer, visible light with a specific wavelength to excite the photosensitizer, and tissue oxygen. Reactive oxygen species that induce antimicrobial effects by oxidative damage are generated after absorption of a suitable wavelength of light by a photosensitizer in the presence of tissue oxygen [11][12][13]. MAL is an ester of 5-aminolevulinic acid (5-ALA) which can be found in nearly all bacterial and human cells as a component of the heme biosynthesis pathway. Accumulation of photoactive porphyrins in targeted cells can  be achieved by providing it exogenously [9]. Cells with a high metabolic rate such as neoplastic cells and bacteria will accumulate 5-ALA more intensely [14]. 5-ALA has already been used in clinical studies, specifically for PDT treatment of C. acnes bacterial biofilms in sebaceous glands [15][16][17][18]. Using MAL-PDT in our study, we aimed to reach not only superficial skin-colonizing bacteria, but also eccrine glands acting as a reservoir of bacteria [15,19,20] that are not affected by topical skin antisepsis treatment. PDT treatment was performed at the Division of Dermatology of the University Hospital of Zurich. MAL (Metvix®, Galderma SA, Tube 2 g) was applied in a sheer layer over the right groin area, covered by a light-impermeable bandage, and allowed to incubate before PDT for 3 h. The area was then exposed to a 633 nm diode (LED; Healite II®; Lutronic; Billerica, MA, USA) for 13 min with an energy dose of 40 J/ cm 2 . Photographic documentation of treated skin was made before and after PDT.

Skin antisepsis with povidone-iodine/alcohol
Antisepsis treatment of the groin was performed 3 times according to standard operation room procedure using Betaseptic (povidone-iodine/alcohol) with at least one-minute interval in between placement.

Biopsies for dermatohistopathology
In four participants, 4 mm skin punch biopsies of the right groin were taken directly after PDT (n = 2) or one week after PDT (n = 2), and again after 3 weeks (n = 4). As a control sample, a similar punch biopsy was taken from the left untreated groin in the same participants (n = 4). Skin biopsies were fixed in formalin and embedded in paraffin. Sections were sliced 4 μm thick for hematoxylin and eosin staining and assessed for the presence of inflammation in the perivascular compartment, focusing on areas with eccrine sweat glands and sebaceous glands. In those specimens with inflammation, the type and amount of inflammatory cells (lymphocytes, neutrophils, eosinophils) were assessed.

Microbiology
Skin swabs were evaluated for any bacterial growth using Columbia sheep blood agar plates without antibiotics (bioMérieux, Mary-l'Etoile, France) and colistin-nalidixic acid blood agar (bioMérieux, Maryl'Etoile, France) plates for aerobic cultivation. Brucella agar plates (inhouse sheep blood agar plates with hemin added Vitamin K1 provided by the Institute of Clinical Microbiology, Zurich, Switzerland) were used for anaerobic cultivation with GENbags (bioMérieux, Maryl'Etoile, France). Cultures were left for at least 7 days at 37°C to check for persisters. Each suspicious bacterial colony was analyzed for bacterial identification using matrix assisted laser desorption ionizationtime of flight mass spectrometry (MALDI-TOF MS).

Results
Half of our included healthy participants in the study (5 out of 10) receiving PDT treatment were female with a median age of 30 years (range 21-54 years) and a median BMI of 22.3 kg/m 2 (19.0-24.9 kg/ m 2 ). Demographic characteristics were similar in the control group without PDT [(7 female, median age of 30 years (range, 22-50 years), and a median BMI of 25.02 kg/m 2 (range, 18.6-26.8 kg/m 2 )].
The effect of MAL-PDT on skin antisepsis did not last 1 or 3 weeks post-treatment, with skin sterility after antisepsis decreasing to 62.2% and 40%, respectively. This was similar to the control group, where we obtained sterility in 50% by performing skin antisepsis treatment only (Table 1). Almost all tested bacterial species recolonized within 7-21 days in varying amounts, with C. acnes presenting most frequently. In addition, Corynebacterium tuberculostearicum was immediately detected after 7 and 21 days. Skin antisepsis reduced the amount of all recolonized bacteria but did not eradicate them completely (Fig. 2B-C). We did not perform statistical analysis due to low patient numbers and did not investigate for the presence of antibiotic resistance.

Histopathology
In skin biopsies, all specimens showed slight inflammation with lymphocytes and eosinophils presenting immediately after PDT on day 0 in the perivascular compartment (n = 2) (Table 2, Fig. 3). There was no inflammation around eccrine sweat glands or around sebaceous glands. After 7 days, similar lymphocytic and eosinophilic presence was detected in the perivascular compartment in one of two participants; this participant had also inflammation with lymphocytes and eosinophils around eccrine sweat glands. Aside from inflammatory infiltrate, Table 1 Skin sterility (defined as no viable bacteria using conventional cultures) before and after PDT treatment and following skin antisepsis treatment.

Sterlity
Nr/Nr total (%) Day 0, control group without PDT, n = 10 Before antisepsis (Skin swab 1) 0/10 (0) After antisepsis (Skin swab 3 Table 2 Histological assessment of skin inflammation. no changes in the histomorphology of sweat glands or sebaceous glands could be observed. After 21 days, inflammation had completely disappeared in all participants (n = 4). The untreated control biopsies from the left groin showed no inflammation throughout the treatment (n = 4).

Side effects of PDT
PDT was generally well tolerated. Side effects (localized redness) were slight and temporary. All participants reported slight redness (Grade 1 according CTCAE V5) during and after PDT, which resolved within 7 days after the treatment (Table 3, Supplementary Figure S1). Seven out of 10 participants reported a slight sensation of burning (Grade 1-2 according CTCAE V5) during PDT, which was comforted with a cooling water spray during the treatment. Itching was only reported by 3 participants during the follow-up period but not during PDT (Table 3).

Discussion
We demonstrated that MAL-PDT in combination with standard skin antisepsis treatment completely inhibited bacterial growth of skin-colonizing bacteria in the groin area of 10 healthy participants. Skin antisepsis or PDT monotherapy achieved less sterility compared to combination therapy with skin antisepsis. However, the effect of PDT on skin antisepsis did not last until 7 or 21 days. Recolonization was represented mostly by S. epidermidis and C. acnes, common skin commensals [21] associated with PJI [1]. Based on these results, MAL-PDT would be a promising tool when used on the day of surgery to eliminate the skin microbiome. However, using PDT with topical MAL cream led to transient skin erythema, possibly from the MAL prodrug that induces the photosensitizing agent Pp IX, which is an obstacle for immediate surgery.
Achieving 100% skin sterility with MAL-PDT and skin antisepsis treatment is an extremely promising study outcome compared to the 50% sterility witnessed using skin antisepsis treatment alone. Bryld and Jemec investigated if using PDT with MAL as a prodrug for the photosensitizer Pp IX could change the normal skin flora in patients undergoing standard PDT for skin disease treatment but reported no significant effects [22]. Compared to 5-ALA, MAL has been shown to have a more selective and deeper distribution of photoactive porphyrins in thick basal cell carcinomas [23] and sebaceous glands [19], with less systemic uptake and pain during irradiation [24]. We chose MAL (ester of 5-ALA) in combination with a red-light source at a wavelength of 633 nm to penetrate deeper into the skin layers and target eccrine glands. Skin biopsies from our participants only demonstrated slight perivascular inflammation immediately after PDT but no changes to eccrine or sebaceous glands [25]. Previous investigation by Itoh et al. described a slight decrease of nuclei of sebaceous glands. Repeated PDT sessions may therefore be necessary to successfully attack all sebaceous or eccrine glands [26].
We documented some mild side effects (local redness, burning sensation) during and after irradiation, resolving within 1 week, coordinating with previous literature [19,25,27,28]. In patients treated for acne, lesions returned to normal skin tone without hyperpigmentation after 1 month [25]. The skin redness after irradiation might be a limiting factor for surgeons concerned with proper wound  healing. However, our histological analysis of the dermis showed only a slight immediate or delayed inflammation, favoring the use of PDT and does not exclude it as a potential prevention strategy. Hence, more studies are needed to determine the optimal time point for PDT prior to surgery to achieve the maximum bactericidal effect with an acceptable local skin reaction. MAL-PDT with less irradiance intensity, reduced duration of the treatment, or other photosensitizers could be tested to minimize local erythema while maintaining the same bactericidal effect. Alternatively, a skin cream with human growth factors and cytokines has been discussed to treat phototoxicity after PDT [29]. Recently, PDT using daylight was presented as a pain free approach to treat actinic keratosis [30]. This could be advantageous over artificial light and decrease skin irritation. Our evaluation for skin commensals was through conventional bacterial culture derived from skin swabs. This took us at a minimum 7 days until we were able to observe and identify colonies. Molecular techniques as real time PCR could facilitate the detection directly from tissue samples since its multiplexing capability allows rapid amplification of several bacteria simultaneously [31][32][33]. However, molecular techniques cannot distinguish between viable and non-viable bacteria, requiring the need for live culture techniques.
This pilot study is the first that examines the effect of PDT on the skin microbiome as an improved implant-associated infection prevention strategy in arthroplasty surgery. The ability to achieve 100% sterility immediately after MAL-PDT and skin antisepsis treatment is striking. However, it should be noted that our study population was small and participants were younger than the average age of patients undergoing arthroplasty surgery, hence larger studies within orthopedic wards are needed to further evaluate our study results. Since different approaches to improve skin antisepsis have been performed so far with no promising effect [4] or may lead to emergence of bacterial resistance [4,8,28], we are convinced that a strategy with an improved skin antisepsis with PDT can provide superior results to decrease PJI.
To conclude, MAL-PDT has an immediate bactericidal effect on the skin microbiome in the groin area. Due to local side effects, more clinical studies with modulation of MAL-PDT, such as decreased light exposure, shorter irradiation time, or other photosensitizers, are needed.

Funding
This work was supported by the Vontobel foundation, Swiss life Jubiläumsstiftung, and Galderma SA, Switzerland who provided the photosensitizer for all participants.

Declaration of Competing Interest
No authors have any conflict of interests.