Adverse Effects of Topical Photodynamic Therapy

A Consensus Review and Approach to Management

S.H. Ibbotson; T.H. Wong; C.A. Morton; N.J. Collier; A. Haylett; K.E. McKenna; R. Mallipeddi; H. Moseley; L.E. Rhodes; D.C. Seukeran; K.A. Ward; M.F. Mohd Mustapa; L.S. Exton


The British Journal of Dermatology. 2019;180(4):715-729. 

In This Article


Characteristics and Frequency

PDT exerts its effects through a phototoxic mechanism; as part of this, pain and inflammation occur. With some of the more conventional higher irradiance topical PDT regimens, pain during irradiation is almost invariable. The mechanisms of PDT-induced pain are poorly understood, but studies in an adenocarcinoma cell line in vitro demonstrated preferential uptake of 5-aminolaevulinic acid (ALA) by beta-amino acid and γ-aminobutyric acid transporters, which was not seen with methyl aminolaevulinate (MAL); this may be one possible explanation for the neurogenic nature of the pain experienced during ALA-PDT, although this was in a cell line model and has not been substantiated in humans.[2] In contrast, MAL uptake has been shown, in a human colon adenocarcinoma cell line, to be mediated by active transport mechanisms involving nonpolar amino acids, providing a potential rationale for any differences in pain mechanisms and experience during PDT following photosensitization by either ALA or MAL.[3]

While there have only been limited studies of the mechanisms of PDT in human skin, it is clear that there is oxidative stress and generation of reactive oxygen species, and an inflammatory reaction involving the release of histamine, nitric oxide, prostaglandin E2, tumour necrosis factor-α and other cytokines; these may also be implicated in the pain and discomfort during and after PDT.[2,4–6] In addition, a neurogenic mechanism involving transient receptor potential (TRP) channels has been implicated.[7–9] A recent study also showed mechanistic differences between ALA and MAL, in that ALA-PDT appeared to induce pain via singlet oxygen-mediated lipid peroxidation, in turn triggering nociceptor activation via TRP cation channel subfamily V member 1 (TRPV1) receptors in dorsal root ganglia in vitro. Furthermore, the TRPV1 inhibitor menthol reduced action potentials evoked by ALA-PDT in dorsal root ganglia and pain behaviour in a mouse model, although this was not the case with MAL-PDT.[10]

In humans, PDT-induced pain begins almost immediately after irradiation starts. Commonly, patients describe a prickling, stinging, sharp burning sensation, most similar to that reported by patients with erythropoietic protoporphyria.[11,12] There is large interindividual variation in the degree and nature of PDT-induced pain experienced by patients, although approximately 16–20% will report severe pain with conventional PDT.[13–16] The multifactorial nature of PDT-induced pain and relative limitations of effective treatment options are well described.[17] In one study that looked retrospectively at experience related to almost 1000 PDT treatments, 44% of patients required some form of pain-reducing intervention.[18] Indeed, in two separate studies – one a survey of PDT services in Scotland and the other a prospective cohort study – of patients treated with PDT for superficial basal cell carcinoma (BCC), squamous cell carcinoma (SCC) in situ or actinic keratosis (AK), 28–38% of patients reported moderate-to-severe pain (score > 6 on a 0–10 numerical rating scale).[19,20] Most of these data are derived from conventional topical PDT regimens using hospital-based, relatively higher irradiance light delivery. However, the PDT procedure is generally very well tolerated, with the pain in the majority of cases resolving once the irradiation period ends [7–9 min with the most widely used red light-emitting diode (LED) sources] and this is reflected in patient preference for PDT over alternative treatments. Nevertheless, the potential for this degree of pain is not ideal for patient care; thus, information on predictive factors and suitable methods of pain relief are required.

Predictive Factors of Photodynamic Therapy-induced Pain

Patient, lesion and treatment site characteristics. The literature relating to possible predictors of PDT-induced pain is complicated by the fact that many of the studies reported are retrospective and have multiple confounding factors. There are conflicting reports of an impact of sex and skin phototype, but there is no clear pattern emerging to suggest a strong effect of age, sex or skin phototype on likelihood of severe pain experienced with topical PDT.[13,18,20–24] More consistently, there is evidence to support PDT to larger treatment areas being associated with more pain;[13,14,16,18,22,23,25] therefore, this has the potential to limit the size of field that can be treated with conventional PDT,[26] although the increasing use of daylight PDT (dPDT) has been beneficial in this regard. Any possible influence of diagnosis and body site is not clear, again owing to potential confounders as, for example, AK tend to affect larger areas and arise on the head and neck. However, reports of PDT used for AK when compared with BCC,[27] acne,[28,29] psoriasis[30,31] and viral warts[32,33] indicate that higher PDT-induced pain scores may be observed when treating these conditions. Thus, it is important to have an awareness of this to minimize any potential impact on treatment delivery and to ensure that patients are appropriately advised and managed.

One study also indicated that there was an association between more severe pain and the degree of erythema in the pretreated lesion.[14] However, this association has not been found by other investigators.[21] Likewise, while Lindeburg et al. reported that the second treatment was more painful than the first in 38 patients treated with PDT,[34] it has, again, not been confirmed by other investigators.[16,21,35] The study of Sandberg et al. also showed that lesions that responded best to PDT were associated with more pain,[14] and it may be intuitive to consider that the more photosensitizer uptake and the greater lesional fluorescence and subsequent phototoxic insult might well lead to the best therapeutic outcome. However, this is not the case when treating AK on the dorsal hands with PDT, as increasing protoporphyrin IX (PpIX) accumulation does not improve efficacy of treatment but increases adverse effects.[36] Subgroup analysis of the larger, multicentre, randomized controlled trials (RCTs) of efficacy of PDT, particularly in dysplasia and superficial nonmelanoma skin cancer, have not been undertaken to investigate whether there is an association between fluorescence, phototoxic inflammation and subsequent therapeutic outcome. Certainly, there is some evidence in smaller studies of a correlation between the degree of fluorescence intensity and pain experienced during PDT,[37] and this has been shown in acne vulgaris and in AK.[28,38] In the latter study, the association with pain was shown between both the degree of PpIX fluorescence and the fluence rate of light delivery, and this is supported by other investigators.[16] Furthermore, pain is not required for PDT efficacy, as exemplified by dPDT, which is considered to be owing to the lower irradiance of daylight and of low level of continuous photoactivation of PpIX.[39]

The influence of prodrug on photodynamic therapy-induced pain. In a double-blind, RCT investigating forearm sites that had been tape-stripped in healthy volunteers, pain was higher on sites exposed to ALA than MAL. In addition, ALA induced higher levels of fluorescence, and there was a greater decrease in fluorescence with irradiation.[40] In a separate study, the same group compared the pain associated with MAL-PDT with ALA-PDT for acne and AK, and showed that the pain experienced was greater with more intense PpIX fluorescence and with a higher rate of light delivery.[38] This greater level of PpIX accumulation and fluorescence associated with ALA has consistently been reported, both in normal skin[40–43] and diseased skin.[28] In addition to higher levels of phototoxicity occurring in normal skin following ALA-PDT compared with MAL-PDT, more prolonged hyperpigmentation may also occur with the former.[43] However, when analysing studies in which MAL-PDT and ALA-PDT have been compared directly, usually there have been other variables, in particular the duration of application of the prodrugs.[44,45] Indeed, two small studies comparing ALA-PDT and MAL-PDT when used for nodular BCC and acne with application for 3 h in each, showed no differences in acute pain scores between the prodrugs,[28,46] although there was greater pain associated with ALA-PDT at 24 h post-treatment in the acne study.[28] More recently, in a large, multicentre study comparing ALA in nanocolloid emulsion (BF-200 ALA) with MAL-PDT, there was no significant difference in adverse effects seen between the prodrug treatment arms.[47] Reduction of drug concentration and/or incubation time may also be considered for effective, less painful PDT, as may be employed for AK or acne.[48–50] With the development of newer formulations of topical prodrugs and lower drug dose regimens, vigilance is required to ascertain whether any change in depth of effect and efficacy may also be associated with changes in pain experienced and tolerance of treatment.[51–53]

The influence of light delivery on photodynamic therapy-induced pain. Most topical PDT is undertaken using LED light delivery. There are few studies in which laser light delivery has been compared with noncoherent broadband light sources, although the evidence from two studies, one of which was retrospective, indicated no significant difference in efficacy or adverse effects, which included pain.[11,54] Certainly, in vitro and in vivo studies support the safety profile of LED light delivery.[55,56] Čarija et al. undertook a within-patient, prospective, controlled study of LED PDT with pulse-dye laser PDT in 15 patients with 62 BCC lesions.[57] While there were similar pain scores between the treatment arms, lower clearance rates were seen at 12 months with pulse dye laser PDT.

In the large, multicentre study comparing BF-200 ALA-PDT and MAL-PDT for AK,[47] more adverse effects were observed in patients treated with a narrower spectrum LED source than in those treated with a broader spectrum, albeit without longer-term safety concerns.[58] Investigators have shown that variable pulsing of light delivery may reduce the pain associated with MAL-PDT for AK in a prospective, controlled study that also showed no loss of efficacy or change in patient satisfaction.[59] Other variables of light delivery have been studied, including the use of filtering of infrared in one study of 80 participants, which was associated with less pain than conventional LED PDT, without loss of efficacy.[60]

Most dermatological PDT uses red light for delivery of depth of effect, but the wavelengths included do impact on PDT-induced pain. In one AK study where green and red light PDT were compared, less pain was experienced using the former with no loss of efficacy in this superficial indication.[61] However, a similar study comparing green and red light for SCC in situ showed loss of efficacy with green light and no significant difference in pain.[62] Mikolajewska et al. undertook a study in 10 healthy volunteers exposed to topical ALA and MAL for 24 h and irradiation was undertaken using either violet laser light or red laser light.[63] In this study, greater pain was experienced in association with red light and a more persistent erythema seen for ALA-PDT, although these differences were not seen in the sites treated with MAL-PDT. However, the results have not been followed up with investigations in diseased tissue and the relevance of this in the clinical setting is unclear. There does not seem to be a strong association between pain experienced during PDT and total light dose used, and this likely reflects the fact that pain is maximally experienced in the first half of irradiation.[64–66] Thus, simply reducing the total dose used is unlikely to impact significantly on the tolerance of treatment.

However, there is substantial evidence that lower irradiance light delivery during PDT, such as dPDT or reduced irradiance hospital or portable device light delivery, is at least as effective as conventional higher irradiance regimens.[38,39,67–73] It seems that at lower irradiances, particularly < 50 mW cm−2, less pain is experienced during PDT.[16,38,39,67–77]

In particular, the use of dPDT has been compared with conventional PDT in large, within-patient, multicentre studies, most recently in Europe and Australia involving patients with mild-to-moderate field-change AK.[78,79] An overall consensus indicates that dPDT to large areas of AK is extremely well tolerated, with much lower pain scores than for conventional PDT, and that efficacy rates are similar.[71] In addition, in support of the use of low irradiance PDT, preliminary data obtained from non-comparative, open studies of low irradiance portable ambulatory LED devices indicate that pain scores are also very low and efficacy at 1-year follow-up is high.[67,72,73,80]

These are important developments for the use of PDT in situations where pain previously could have been a treatment-limiting factor. This now enables larger areas to be treated in a well-tolerated and an almost painless, effective regimen with dPDT. Another alternative means of varying irradiance using conventional hospital-based LED devices is with use of an initially reduced irradiance at < 50 mW cm−2, and, thereafter – for the latter part of the regimen – to increase irradiance in order to deliver an overall effective light dose. This approach of increasing irradiance during PDT after an initial lower (< 50 mW cm−2) irradiance approach to light delivery may be associated with reduced pain scores and can be useful, for example, if treating genital or perineal sites.[81] This was investigated in a retrospective, single-arm study of 14 patients treated with this two-step irradiance regimen for BCC and SCC in situ, showing high clearance rates.[82,83] Fractionation of light has also been investigated as a means of improving efficacy of PDT,[84] although this has been shown to be at the expense of increased adverse effects, notably pain.

Pain: how does photodynamic therapy compare with other treatments?. When looking at the outcome of severe pain which requires a break in treatment or use of local infiltration anaesthesia, PDT results in significantly higher pain scores compared with placebo.[29,47,85,86] This is also the case for lower levels of more manageable pain. Furthermore, when comparing dPDT with conventional PDT, the former is significantly less painful, based on large, multicentre studies.[39,71,74,78,79]

In the larger studies comparing the outcome of severe pain, which requires a break in treatment or use of local infiltration anaesthesia, no significant differences were seen between cryotherapy, 5-fluorouracil or imiquimod,[35,87–89] whereas less pain was experienced with surgical excision than with PDT, although this would be expected as local anaesthesia is used for the surgical procedure.[90,91] Of note, the pain and discomfort of other topical treatments, such as 5-fluorouracil or imiquimod, is not directly comparable with PDT; the former are associated with increasing discomfort and inflammation during the course of treatment, over several weeks, whereas the pain experienced by PDT is maximal in the first few minutes of treatment and then subsides rapidly.[35,92] This is an acute, rather than a more chronic experience, probably indicating why patient satisfaction levels with PDT are high.[92]

Thus, when MAL-PDT was compared with ingenol mebutate for treatment of multiple AKs on the face and scalp in within-patient, split-face studies, pain scores and cosmetic outcome were higher with PDT, but local skin reactions were more severe and persistent with ingenol mebutate; overall, patients preferred PDT.[93,94] When dPDT was compared with ingenol mebutate in 27 participants with AK in a within-patient study, pain scores were higher for ingenol mebutate.[95] Similar efficacy was reported between the two groups but increased tolerance for dPDT was documented in terms of reduced local skin reactions and pain, and preference for dPDT.[95]

Furthermore, in a randomized, observer-blinded, within-patient comparison of patients with multiple AKs treated with trichloroacetic acid compared with ALA-PDT, higher pain scores and efficacy rates were seen with PDT and scarring was present only in those treated with trichloroacetic acid.[96]

Pain Relief for Photodynamic Therapy-induced Pain

Treating with methods of no significant benefit. Given the nature of PDT-induced pain and the probable neurogenic mechanisms involved, it may be anticipated that topical anaesthesia could be beneficial for pain relief during PDT. However, in a within-patient, double-blind RCT of ALA for extensive AK on the scalp, Langan and Collins failed to show a significant effect of eutectic mixture of local anaesthetics (EMLA) for PDT-induced pain.[97] This is supported by the observations of Grapengiesser et al. in 60 patients in whom EMLA was used during PDT.[13] A separate interindividual study by Holmes et al. found no significant effect of tetracaine gel (Ametop®; Perstorp Speciality Chemicals AB, Perstorp, Sweden) used topically during ALA-PDT for superficial BCC, SCC in situ or AK.[98] Likewise, during large-area PDT for facial AK and field-change carcinogenesis, no benefit of topical lidocaine hydrochloride 3% cream was found.[99] In a randomized, double-blind, placebo-controlled study, morphine gel 0·3% was shown not to be significantly beneficial for pain relief during topical MAL-PDT;[100] Sandberg et al. observed that capsaicin cream was also not significantly effective in reducing pain and there were side-effects of the topical capsaicin itself.[14]

Treating with methods of potential benefit. In contrast, a pilot, open, split-face study performed by Borelli et al. on the use of subcutaneous infiltration of ropivacaine 1% with prilocaine 1% for PDT pain relief showed benefit, although there were significant adverse effects of cheek swelling persisting for up to 3 days, which could limit its use.[101] This has been supported in a separate case report showing the benefit of subcutaneous anaesthesia for pain relief during PDT in a 7-year-old child.[102]

In addition, peripheral nerve blockade can be significantly effective in reducing PDT-induced pain when used for extensive facial AK. In an initial study in 16 patients with symmetrical facial AK, nerve blockade using mepivacaine and adrenaline was used to block supraorbital, supratrochlear, infraorbital and mental nerves, and the nonanesthetized side served as control. Pain scores were significantly reduced on the anaesthetized side and 15 of the 16 patients expressed preference for nerve blockade in future if PDT was required.[103] This has also been supported by a separate study in 10 male patients with facial AK, using supraorbital, supratrochlear and occipital nerve blockade during MAL-PDT.[104] In an open clinical trial involving 34 patients with frontal facial AK where supraorbital and supratrochlear nerve blockade was used on one side and cold air analgesia on the other, nerve blockade was significantly superior with respect to pain relief, and preferred in 31 of the 34 patients.[105] However, nerve blockade is only possible at certain body sites, and, of course, requires an additional invasive procedure; as such, it may not be appropriate for many patients treated with PDT.

In a prospective, controlled, observational study to address the potential effect of nitrous oxide (NO), involving 71 patients treated with MAL-PDT for multiple AKs on the cheeks, all patients received 800 mg ibuprofen 30 min before PDT irradiation. In addition, cooling was used with a cold-air fan and interruptions in treatment were allowed if required and, for patients who experienced severe pain (visual analogue scale score ≥ 6), despite ibuprofen and cooling air, additional NO and oxygen mixture (Livopan®) was offered for PDT to the other cheek. Overall, a reduction in pain score of 55·2% was seen between treatments to the first and second cheek following application of the NO and oxygen mixture. Treatment was generally well-tolerated, although six of 30 patients (20%) experienced mild side-effects during inhalation of the NO and oxygen mix, which included vertigo, fear of loss of control and amplification of noise.[106]

Considering other options for pain relief during PDT, investigators have explored the potential use of transcutaneous electrical nerve stimulation (TENS). This was undertaken in a pilot study in 14 patients with facial and scalp AK who had experienced severe pain during earlier PDT treatments. When the TENS electrodes were placed on the shoulders, four patients found no benefit from the use of TENS and three patients (21%) who had had previous interrupted PDT sessions due to pain were able to complete treatment, although the reduction in pain scores was modest (from 8·1 to 6·2). Overall, all but one patient would have used TENS again during PDT. This pilot study requires further investigation, although TENS is only feasible at certain body sites and therefore may have limited application in routine clinical use.[107]

In many PDT regimens, use of a cold-water spray is employed as a routine measure during PDT. In a double-blind, controlled study involving 85 patients treated with ALA-PDT for AK or acne vulgaris, two thermal spring waters were investigated and sprayed four times daily to the face for 1 week after PDT. A reduction in discomfort, pain and erythema was experienced between days 2 and 7, although no impact was shown on the period of maximal pain, which was on day 1.[108] In a separate study of 24 patients with AK treated with MAL-PDT on two symmetrical areas, cooling with either cold-water spray or a cold-water pack was employed in either the first or second period of illumination. The water spray and cool pack reduced mean pain scores modestly by 1·2–1·3 points; however, pausing irradiation was associated with a higher reduction in pain of 3–3·7 points. Thus, while cooling resulted in minor reduction in pain intensity, a pause in illumination was more effective for pain relief, and these are relatively easy things that can be incorporated routinely into clinical PDT practice.[109] Pausing during illumination may also be useful when treating acne with PDT.[51] The relatively small impact of cooling air on reduced PDT-associated pain was also shown by Stangeland and Kroon, who undertook an open, within-patient, right–left comparison study in 43 patients treated with MAL-PDT for field change cancerization, showing a small but significant reduction in pain scores in those treated with cold air analgesia.[110] These observations of the utility of cooling are supported by a nonrandomized, retrospective, observational, controlled study in which cooling devices were seen to be associated with reduced PpIX photobleaching. However, a reduction in disease clearance rate was seen at 3 months of follow-up and thus cooling should be used with caution because of concerns about adverse impact on therapeutic effect.[111]

Other treatment methods. Less conventional approaches have included a plant-derived spray that contained camomile and menthol, which was used in addition to glycolic acid. A randomized, blinded study involving 56 patients with field-change cancerization of either arm (n = 25) or face (n = 31) showed reduced pain scores at all time points up to 30 min, during and after treatment. The sprays were applied to treatment areas 10 min before irradiation and at any time during irradiation, with the placebo being a coffee-scented saline spray.[112] While this may be a relatively simple method to reduce discomfort when large areas are treated with PDT, it needs further study.

While PDT is generally well tolerated, exploring options for patients who have found PDT to be painful is worthwhile. A single session of hypnosis was explored in a pilot study of 12 patients treated with PDT for precancerous lesions (actinic keratosis, SCC in situ, Bowenoid papulosis and Paget disease), showing significantly reduced pain scores in eight patients, six of whom had previously experienced PDT without hypnosis. While it would not be required for most patients treated with PDT, hypnosis requires further investigation as it could be considered in exceptional circumstances if proven to be effective. A limitation would be the requirement for members of staff to be trained adequately in hypnosis.[113]

Thus, while nerve block, subcutaneous infiltration with anaesthetic, TENS, cooling air and/or pausing irradiation may be of benefit, more typical forms of topical anaesthetics or oral analgesics have not been shown to be effective.[20,114] Modifying PDT regimens to employ lower irradiance light delivery is usually most effective, enabling successful treatment.[7,115]