4.1 The Role of UVR in Photodistributed SJS/TEN
The nature of the rash in these cases provides evidence that UVR precipitates SJS/TEN in patients taking susceptible drugs. Rash timing occurred after sun exposure and originated on sun-exposed areas of the body. As the rash progressed over time to form vesiculobullous lesions with skin sloughing, the photodistributed pattern remained. It is the photodistributed pattern that led the authors to name these disorders photodistributed SJS, photodistributed SJS/TEN overlap, and photodistributed TEN, referred to collectively in this study as photodistributed SJS/TEN. Furthermore, a unique finding in this study is that no cases describe a flu-like prodrome, which generally proceeds rash onset by approximately 3 days in SJS/TEN [1,41]. Photo-induced cases generally describe the rash as their primary manifesting symptom of their disorder. As traditional cases of SJS/TEN do not manifest initially with a rash, but a flu-like prodrome, and do not present in a photodistributed pattern, this is evidence that UVR precipitates the disease process, altering its course and clinical presentation.
Further evaluation of the rash reveals that the photodistributed rash presented after a minimum of 24 hours from UVR exposure with non-discrete borders (less demarcated than a sunburn) and satellite lesions on UVR protected (clothing covered) skin (Fig. 2). Furthermore, the rash commonly spreads to the palmar surfaces of the hands and the plantar surfaces of the feet, areas that are reported as receiving less UVR exposure compared to other areas of the body [42,43]. The delayed onset of the rash, with extension onto areas with less UVR exposure, along with the observed finding of inflammation and ulcerations of mucous membranes, supports that UVR triggered a systemic immune response that is consistent with SJS/TEN, even though sun-exposed areas remained the most severely affected.
Additional evidence that elucidates the role of UVR in SJS/TEN cases was presented in a Chilean study after an analysis was performed on 24,521,796 hospital discharges nationwide [44]. The study revealed that the incidence of SJS/TEN occurred with an increased frequency in regions of the country with higher altitude, leading to the conclusion in the study that the additional UVR exposure at high altitudes may play a role. The photodistributed rash, the altered primary manifesting symptoms (rash vs. flu-like prodrome), the evidence of a systemic immune response that remains most prominent in sun exposed areas, and the increased incidence of SJS/TEN in areas of increased UVR exposure all suggest that UVR precipitates the onset and clinical presentation of photodistributed SJS/TEN cases.
4.2 Causal Drug
All cases that were included in this review reported a causal drug, even though the search criteria did not require a causal drug. Additionally, no cases reported that UVR caused SJS/TEN by itself. Therefore, it is likely that the drug plays a key role in the pathogenesis of photodistributed SJS/TEN. It is postulated that the direct effect of UVR on the drug itself, or the surrounding cutaneous tissue that then affects the drug, leads to the sequelae of the SJS/TEN presentation (see 4.3 Pathogenesis Theories of Photodistributed SJS/TEN). Thus, it is important to validate that the reported drugs were responsible for their cases of photodistributed SJS/TEN. Drug causality assessment tools (CAT) are limited in their ability to determine culprit drugs in SJS/TEN cases. Their findings have been shown to vary between investigators [45]. Consequently, the results should always be compared to a thorough clinical assessment. If a drug CAT is to be employed in a SJS/TEN case, the ALDEN CAT is arguably the most reliable [45]. In this analysis, all cases, apart from one with insufficient information, revealed that the identified drugs were very probable, probable, or possible for causing SJS/TEN. These results, along with a complete review of each case, revealed that the identified drugs could have reasonably acted as the culprit drugs.
4.3 Pathogenesis Theories of Photodistributed SJS/TEN
Due to technological advancement and continued research, our understanding of the pathogenesis of SJS-TEN continues to be better understood. Drugs are implicated as the causative agent in as many as 80% of cases [46]. The sensitization of the immune system to a causative drug has been described as a T-cell mediated immune response, causing a delayed type VI hypersensitivity reaction. This type of response requires the patient to have a specific HLA allele that codes for a major histocompatibility complex (MHC) that is specific for a drug or drug-metabolite (antigen) [47]. Therefore, certain drug-induced cases of SJS/TEN have been linked to specific HLA haplotypes [47]. Recognition of the MHC-antigen complex by specific T cell receptors results in T cell activation (sensitization). The exact sequela of immune sensitization in SJS/TEN is still being investigated. Histopathological outcomes from the immune response include subepidermal blisters with widespread necrosis and apoptotic keratinocytes associated with minimal lymphocytic inflammatory infiltrate [1]. It has been observed that there is high expression of granzyme, perforin, and FasL in mononuclear cells in TEN blisters, suggesting CD8+ T-cell activation which is likely responsible for the epidermal keratinocyte necrosis [48,49].
UVR plays a key role in the pathogenesis of photodistributed SJS/TEN. There are several classifications of UVR based off wavelength: ultraviolet A (UVA) = 320-400 nm, ultraviolet B (UVB) = 290-320 nm, and ultraviolet C (UVC) = 200-290 nm [50,51]. It is generally accepted that the majority, if not all, of UVC radiation does not make it through the atmosphere to reach the surface of the earth [50]; however, UVB and UVA radiation do make it to the earth surface and therefore affect human skin. UVB radiation is readily absorbed by the DNA in the epidermal keratinocytes and as deep as the papillary dermis [52]. UVA radiation is minimally absorbed by the DNA of superficial keratinocytes, but penetrates deep into the reticular dermis where it is absorbed by other molecules and proteins, including drugs that have distributed to the area [52,53]. More precisely, UVR is electromagnetic radiation that travels from the sun in waves containing photons [52]. These photons are composed of high energy, which are absorbed in the cutaneous layers, that induces electrons to move from their ground state to an unfilled outer electron shell producing an excited state, a high energy state, that is unstable. This is the process that occurs in drug-induced photosensitive reactions. [52,54] UVA radiation is most commonly responsible for photosensitivity reactions, although UVB may also contribute or be responsible in select cases [55].
Neither type of photosensitivity reaction (phototoxicity and photoallergy) alone can account for the patient cases in this review. However, all of the causal drugs in this review have been implicated in causing photosensitive reactions [52,56–80] (Table 5). Thus, when considering the findings of photodistributed rashes that present after UVR exposure in the setting of drug use (drugs known to cause photosensitive reactions), it is feasible that the mechanisms behind drug-induced photosensitive reactions could, at least in part, contribute to photodistributed SJS/TEN.
A direct drug phototoxicity pathway is one in which UVR induces changes to the drug chemical structure, producing a free radical. The free radicals act on lipids, proteins, and DNA, causing direct damage to cells [81]. An indirect drug phototoxicity pathway is one where the energy from UVR results in reactive oxygen species (ROS). ROS are formed by drug free radicals reacting with oxygen or excited drug states that allow for the transfer of energy to oxygen via excited triplet cascade [81]. An example of ROS formation in response to UVR was observed when a cancer treatment was evaluated. A combined exposure to cells with UVA irradiation and lomefloxacin caused higher alterations of redox signaling pathways, causing intracellular ROS overproduction and endogenous glutathione depletion in melanoma cells [82]. This had an advantageous effect against melanoma cancer cells. However, excess formation of ROS may occur in healthy tissues, exceeding the body's antioxidant defense mechanisms, allowing for oxidative stress to cause damage to cellular components [83]. The cytotoxic effect of ROS in the skin triggers the immune system to attract T-cells to the dermis. In patients with vitiligo, ROS-impaired keratinocytes were shown to mediate CD8+ T-cell infiltration [84]. As CD8+ T-cells have been shown to be the prominent cells in the pathogenesis of SJS/TEN, it is feasible that the cellular damage caused directly by the photoproduct (a photo-induced drug metabolite), or indirectly through the formation of ROS, triggers immune cells to infiltrate the area. Drug and/or drug photoproducts in the skin recognized as foreign may induce drug sensitization, ensuing the immunological sequelae that results in a clinical presentation of SJS/TEN.
In contrast to phototoxicity, photoallergy is immune mediated: a cell-mediated type IV hypersensitivity reaction. When a drug in the cutaneous tissue absorbs the photons, it is converted into a biologically reactive photoproduct. The photoproduct can then act as a complete antigen or a hapten [57]. The complete antigen model is consistent with the p-I concept, proposing that a drug (or photoproduct) itself is sufficient to bind to MHC and TCR to sensitize the immune system [47,85]. In the hapten model, the photoproduct binds to protein within the cutaneous tissue (haptenization) to form a complete antigen [57,86]. Langerhan cells process the antigen and present it to MHC class II molecule and induce the subsequent cell mediated hypersensitivity response, resulting in the homing of activated T-lymphocytes into the skin. As both photoallergy and SJS/TEN reactions are both type IV hypersensitivity reactions that are a result of the immune system responding to a drug, it is reasonable that the immune system could be triggered to go down the SJS/TEN pathway. See 4.4 Differential Diagnoses for more on photosensitive reactions.
An alternate theory is the Koebner phenomenon which describes the appearance of new skin lesions due to trauma, commonly seen in patients with psoriasis and has been postulated as the possible pathophysiology for photodistributed erythema multiforme (PEM) [87]. This theory postulates that UVR causes increased vascular permeability, facilitating the passage of skin antigens into the bloodstream and favoring the formation of circulating antibodies and immune complexes in sun-exposed areas [31,88]. Thus, UVR precipitated koebnerization leads to a photodistributed pattern [38]. The increased passage of skin antigens may allow for increased susceptibility to the drug/drug metabolites/photoproducts, activating the immune response and causing SJS/TEN.
The next theory is the alterations in intercellular adhesion molecule 1 (ICAM-1) expression. UVR has been shown to induce tumor necrosis factor-alpha (TNF-alpha) secretion by keratinocyte, resulting in increased ICAM-1 expression [22,89]. Theoretically ICAM-1 may play many roles in the sensitization of the immune system to the drug/drug-photoproduct in photodistributed SJS/TEN. The expression of ICAM-1 by antigen presenting cells (APC), commonly Langerhans cells in the epidermis [90], lowers the concentration of antigen required to activate a naïve T-cells into effector and memory T-cells [91]. Additionally, ICAM-1 can promote long duration of contact between T-cells and APCs [92]. This means that indirectly, via increased ICAM-1 expression, UVR may ultimately play a role in sensitizing the immune system to a low drug dose, precipitating a reaction of SJS/TEN that otherwise may not have happened. This would help explain why three of the reported cases in this review occurred after one dose of a drug [31,34,39]. Furthermore, when UVR exposure causes a release of TNF-alpha by keratinocytes, it causes an increased ICAM-I expression by keratinocytes in the epidermis and attracts CD8+ T-cells [22,89,93]. CD8+ T-cells show a clear predominance in epidermis in early stages and through various proposed mechanisms trigger epidermal necrosis [1,49]. The increased immune response and cell death in UVR exposed skin could explain the photodistributed presentation in these cases.
The last theoretical mechanism is an increase in Fas/FasL expression. Fas ligand (FasL or CD95L) belongs to the TNF family and is responsible for inducing apoptosis in Fas+ cells [94]. Keratinocyte apoptosis is commonly observed in early stages of TEN [49], consistent with large amounts of lytically active FasL that have been shown to be expressed in SJS/TEN [95]. Monoclonal antibodies that interact with Fas/FasL have shown to block the cytological activity, confirming that FasL expression is responsible for the apoptosis that occurs in SJS/TEN [49]. It has been shown that UVR causes FasL upregulation in the epidermis [96]. FasL may be upregulated due to action of TNF-alpha, released by keratinocytes themselves as a result of UVR [89,97]. Fas receptor expression on keratinocytes has also been shown to increase from both UVB and UVA radiation exposure, with a delay to peak expression at 24 and 12 hours respectively [98]. Upregulation of Fas/FasL in keratinocytes may contribute to the increased involvement of the UVR exposed skin. The activation of the Fas-mediated apoptosis pathway in keratinocytes is presumably activated by toxic drug metabolites [99]. Thus, UVR may increase apoptosis in sun-exposed areas by increasing Fas/FasL expression and creating photoproducts that activate the pathway and further stimulate the immune response in SJS/TEN cases.
4.4 Differential Diagnoses
4.4.1 Photosensitive Reactions
Photosensitivity reactions consist of phototoxic and photoallergic reactions. These reactions are considered the most common type of drug-induced rashes that present in a photo-distributed pattern. As the cases in this study appear to present at least partially like drug-induced photosensitive reactions, it is important to distinguish characteristics of phototoxic and photoallergic reactions from photodistributed SJS/TEN.
Phototoxicity is dose dependent and can happen in anyone who ingests (oral) or applies (topical) enough offending agent with concomitant UVR exposure. These responses are unpredictable between people and even repeat exposures [52]. Clinically, they commonly manifest as a severe sunburn with sharp demarcated lines with or without bullous formation. Phototoxic effects generally result in erythema, appearing immediately within minutes to hours, or less commonly a late onset erythema may occur that takes 24-120 hours to appear [54]. Generally, these reactions reach a peak of maximal clinical manifestation after 24-48 hours of UV exposure [52]. Histopathologically, necrotic keratinocytes are seen along with a predominantly lymphocytic and neutrophilic dermal infiltrate [100].
Photoallergy requires an immunological response that only occurs in sensitized patients. This requires re-exposure of a drug or the continued use of a drug for sufficient time to allow sensitization to occur, reported as three days to years [48]. Most commonly it manifests as an eczematous rash that is less demarcated, often spreading outside of UVR areas. This reaction is not dose dependent; a small amount of drug can trigger the reaction. Symptom onset is generally 24-48 hours after UVR exposure [101]. After symptom onset, the peak of maximal clinical manifestation is around 48-72 hours [52]. Histopathologic features are identical to those seen in an allergic contact dermatitis, including epidermal spongiosis, vesiculation, exocytosis of lymphocytes, and a perivascular inflammatory infiltrate [100,102]. Spongiosis represents the histological hallmark of intercellular epidermal edema, corresponding with wide spaces between keratinocytes and elongated intercellular bridges (“spinous processes”), leading to a sponge-like appearance of the epidermis [103]. The most common cutaneous lesion associated with spongiosis is eczematous dermatitis [104], consistent with the eczematous type of rash seen in photoallergic reactions.
Photodistributed SJS/TEN share more commonalities with photoallergic reactions than phototoxic reactions. Photoallergic reactions require sensitization of the immune system, the onset of the rash is generally > 24 hours from UVA exposure [22,29,30,34–36,39], the incidence is rare, they require a low dose of medication [31,34,39], and the rash distribution may extend past the sun exposed areas [22,28,31–36,38,39], although these areas are generally less-affected than sun-exposed areas. No cases in this review report cross-reactivity of related agents. However, cases of drug-induced SJS/TEN due to drug cross-reactivity have been reported in the literature [105–107]. The similarities of photodistributed SJS/TEN, all shown to be photo-drug-induced SJS/TEN, is likely because drug-induced SJS/TEN occurs due to a delayed type IV hypersensitivity reaction, similar to photoallergic reactions [1,54]. Nevertheless, CD8+ T-cells are more prominent in triggering the immune system in SJS/TEN vs. CD4+ T-cells in photoallergy [108]. The pathophysiology of photodistributed SJS/TEN is likely distinct from other drug-induced SJS/TEN cases as UVR plays a role. Nevertheless, it is assumed that a delayed type IV hypersensitivity reaction ultimately takes place as the cases meet diagnostic findings consistent with SJS/TEN.
Findings not observed in a photoallergic or a phototoxic reaction, but that are commonly found in photodistributed SJS/TEN include mucous membrane involvement [22,28–36,38,39], a positive Nikolsky’s sign [22,30–32,39], rash involvement of the palmar and plantar surfaces [28,30,36,38,39]. There is also a unique timeline of maximal clinical manifestation of the rash, reported as 24-48 hours in phototoxic reactions, 48-72 hours in photoallergic reactions, but commonly progress for days to weeks in photodistributed SJS/TEN [22,28–33,36,38,39]. The histopathological findings are also distinct. Phototoxic reactions often have necrotic keratinocytes, but less commonly full-thickness epidermal necrosis as in SJS/TEN and the immunological findings are distinct in the dermal/epidermal layers and vesicle fluid (if vesicles are present in the phototoxic reaction) [22,34,45,49]. Although DIF findings are rarely reported in photosensitive reactions, DIF has been reported as positive in photoallergy [109] and negative in phototoxicity [110]. DIF is commonly reported in SJS/TEN and is negative. For additional comparisons see Table 6 [100].
4.4.2 Photodistributed Erythema Multiforme
Photodistributed erythema multiforme (PEM) is rarely reported, but an important differential diagnosis. In one review in 2012, 18 cases of PEM were identified: 10 were drug-induced, four were due to herpes simplex virus, one polymorphous light eruption, and three idiopathic [88]. Erythema multiforme (EM) shares many features with SJS and TEN and all were once considered to be in the same spectrum [111]. Further investigation revealed that EM is considered distinct from SJS/TEN. EM is important to distinguish from SJS/TEN as it is self-limiting with mild or no systemic symptoms and usually resolves without complications, versus a systemic reaction that is potentially life threatening in SJS/TEN [2]. EM is a hypersensitivity reaction that may be caused by drugs but is mostly caused by infections such as herpes simplex virus and mycoplasma [112]. Mucous membrane involvement is only associated with 25-60% of EM cases [113]. EM is characterized by a raised, papular rash with “typical target” lesions with three concentric zones; whereas a flat, macular rash with poorly defined “atypical target” lesions in SJS/TEN generally have two zones and demonstrate confluence of lesions [113,114]. In SJS/TEN the rashes progress to have more extensive vesicle/bullae formation and are followed by skin sloughing. Palmar and plantar rash can be seen in EM and SJS/TEN. EM is symmetrically distributed on the distal extremities with minimal epidermal detachment, often to 1% or 2% of BSA (<10% of BSA) [113]. Histologically, the conditions can appear similar in early stages; however, in later stages SJS/TEN can be distinguished from EM: established SJS/TEN shows full-thickness keratinocyte necrosis that develops into subepidermal bullae vs. scattered necrotic keratinocytes that appear in the lower layer of the epidermis (this finding may also be seen in early stages of SJS/TEN) [115]. It has also been noted that the center of a blister in a typical target lesion in EM may also demonstrate full thickness necrosis [116]. Thus, a thorough histological evaluation is warranted. The gravity of the overall clinical presentation, including the involvement of systemic symptoms and the severity of cutaneous and mucosal involvement may suggest SJS/TEN vs. EM early in the patient work-up; however, less severe cases may be less clear and require a detailed work-up to differentiate conditions.
4.4.3 Autoimmune Bullous Diseases, Porphyria Disorders, and SJS/TEN-Like LE
UVR is known to induce or aggravate autoimmune bullous diseases, including pemphigus foliaceus, pemphigus vulgaris, bullous pemphigoid, and less commonly, linear IgA dermatosis [117]. Bullous diseases have many similarities to drug-induced SJS/TEN. They may present with bullae formation, mucous membrane involvement, a positive Nikolsky’s sign and are widely associated with drug use [13,118,119]. Bullous diseases contain autoantibody mediated acantholysis at the basement membrane zone which results in a positive DIF. In contrast to these findings, SJS/TEN is due to necrosis of cells, resulting in the absence of autoantibodies and a negative DIF. Porphyria cutanea tarda and pseudoporphyria are also considered photosensitive bullous diseases that can present in a photodistributed pattern with positive DIF findings that are consistent with the deposition of immunoglobulins and C3 around blood vessels in the dermis and at the dermal epidermal junction [120–122]. Pseudoporphyria is unique in that it is also commonly induced by drugs [122]. Thus, a negative DIF will also help rule out these conditions as potential causes. Furthermore, SJS/TEN-like LE is an important alternate diagnosis that should always be considered, especially in cases that are photodistributed. In addition to a serological autoimmune diagnostic work-up for LE, patients should have a DIF test performed on a biopsy sample of the rash. Patients with SJS/TEN-like LE will have deposition of IgM and IgG in the basement membrane (lupus band) [123]. Additionally, acute cutaneous LE, subacute cutaneous LE, discoid LE, and systemic LE all commonly result in positive DIF of cutaneous samples [124]. In this review, all four cases that reported findings of DIF were negative [22,32,36,38]. DIF should be performed on lesional biopsies in every histopathological workup to help rule out various differential diagnoses that may be photo-drug-induced.
4.5 Distinguishing factors for Photodistributed SJS/TEN
The following criteria was proposed by the authors to help guide physicians when distinguishing photodistributed SJS/TEN from an alternative diagnosis (Table 7). This criteria should continue to be reviewed, and updated when appropriate, with future investigation of photodistributed SJS/TEN.
Must identify = need all 4:
1. A photo-distributed rash as a result of UVR exposure after drug initiation (or other known SJS/TEN cause). The timeline of UVR is important as a recall-phenomenon has been reported which likely represents unique pathophysiology [125,126].
2. Rash progression/worsening that continues for a minimum of 48 hours with rash evolution to include vesicles/bullae and/or sloughing of the epidermis. This is a systemic immune disorder that progresses over days to weeks.
3. Negative DIF. If positive, LE, bullous diseases, and porphyria disorders should be further investigated.
4. More likely SJS/TEN than EM. It is important to make this distinction due to mortality risk in SJS/TEN patients. If the disorder is EM, it is self-limited, does not require specialized treatment in a burn center, and the patient is able to perform additional photobiological testing to aid in culprit drug identification: phototesting followed by photopatch testing [88]. Photobiological testing is too risky in SJS/TEN as re-exposure may induce another SJS/TEN case, resulting in a high risk of mortality [1].
Major criteria (should identify) = need 3/4:
1. A causal drug. All cases in this review include a casual drug, while cases without drug involvement have been reported in SJS/TEN-like LE, PEM, bullous diseases, and porphyria disorders.
2. Mucous membrane involvement. Oral, ocular, and/or genital mucositis is reported in 92-100% of SJS patients and nearly all TEN patients, with oral involvement being the most common [113]. It is suspected that the incidence of mucositis would be similar in photodistributed cases as all 12 cases in this review that reported mucous membrane findings reported mucositis.
3. Positive Nikolsky sign. Nearly all should have, but not a “must identify” criteria due to limitations of the test. (See 4.7 case inclusion consideration for more on Nikolsky sign).
4. Histopathology demonstrating full-thickness epidermal necrosis with subepidermal bullae development. Lesions found in earlier stages of SJS/TEN may not yet have progressed to full-thickness epidermal necrosis, excluding this finding as a “must identify.”
Minor criteria (common findings) = need 2/4:
1. Palmar and plantar rash. No formal report of the incidence could be identified in the literature, although it is commonly reported in SJS/TEN [12] and in the results of this review.
2. Delayed rash onset that presents no earlier than the day after UVR exposure. This was observed in this report and consistent with a delayed hypersensitivity reaction commonly reported in SJS/TEN [1].
3. Satellite lesions/non-discrete borders on UVR protected (clothing covered) skin. These findings were consistent in all cases in this review (those that provided images) to varying degrees.
4. No flu-like prodrome. This finding was consistent with all cases in this review.
Minor criteria 2-4 were important findings in this review that need additional evaluation from further accumulation and investigation of cases to assess their prevalence before they can be considered a “must identify” or “major criteria.” These three findings, in addition to an overall photodistributed rash (number 1 in the “must identify” section), are the findings that are unique from traditional SJS/TEN, summarized as a photodistributed rash with non-discrete borders, occurring at least a day after UVR exposure, and without a flu-like prodrome. The remaining criteria are consistent with traditional SJS/TEN and are important to help distinguish from other photodistributed disorders.
Total % of BSA involvement can help further distinguish between photodistributed SJS, SJS/TEN overlap, and TEN [2]. Although DIF has shown to have significant diagnostic value in LE [123,124,127], the negative predictive value of a negative DIF to rule out SJS/TEN-like LE has not been calculated. Thus, physicians should order lupus autoimmune serology to further contribute to LE rule out. Clinical judgment and appropriate diagnostic workup should be included to investigate differential diagnosis previously discussed.
4.6 Treatment
The authors recommend following up to date treatment guidelines for SJS/TEN, emphasizing the importance of a quick diagnosis and transfer to an appropriate care facility, like a burn center [128]. All patients in this study responded well to their treatments and survived, receiving various combinations of systemic steroids, antibiotics, cyclosporine, intravenous immunoglobulin G, and advanced wound care. Patient education should include avoiding causal drugs and mitigating sun exposure with UVR protection when outside.
4.7 Case Inclusion Consideration
Two cases were less clear about meeting the criteria of disease onset of SJS/TEN after UVR exposure. First, a female with a history of abnormal reactions to ibuprofen chose to take ibuprofen for menstrual cramps which resulted in lip swelling and a minimally reactive rash of small red macules and papules [34]. The rash was stable and non-progressive for two days when the patient decided to self-treat with UVR from a tanning salon that she frequented, triggering the onset of TEN by the following morning [34]. This case was included in the review as the baseline rash was reported to be stable, non-progressive, and radically changed into photodistributed TEN after UVR exposure. The second case was an abstract for a poster. Consequently, it provided fewer details of a female with a history of epilepsy who experienced a case of photodistributed SJS from the ingestion of lamotrigine [37]. This case was also chosen to be included since the photodistribution of the rash implies UVR exposure.
Additionally, cases that reported Nikolsky’s sign findings are worth discussing. A positive Nikolsky’s sign is observed when light pressure is applied to the skin, usually from a clinician’s finger, resulting in the disassociation of the epidermis from the dermis [129]. Five of six cases reported a positive Nikolsky’s sign [22,30–32,39], a common reported finding in cases of SJS/TEN [1,130]. Eloranta et al. [36], presented the only case that reported a negative Nikolsky’s sign. Further evaluation revealed that the histopathologic report of the skin biopsy showed a necrotic epidermis and interface dermatitis with vacuolization and a negative DIF [36], consistent with findings in other cases of SJS/TEN that presented with a positive Nikolsky’s sign [22,30–32]. The Nikolsky’s sign is a clinical test attempting to evaluate if the epidermal cells have detached from the dermis which was suggested in the biopsy findings in this case. Obtaining a positive Nikolsky’s sign depends on correctly performing and interpreting the findings. One case showed that although 75% of a patient’s BSA presented with a rash consistent with TEN, the Nikolsky’s sign was only positive in 9% of BSA upon presentation, which progressed to 20% when re-evaluated 24 hours later [107]. Thus, if a Nikolsky’s sign is only performed on part of the rash, a false negative may result. Additionally, the timing of the test matters. Tests performed later in the disease process are more likely to be positive. This correlated with histological findings that show progression from individual keratinocyte apoptosis to full thickness necrosis in later stages. Due to variability in Nikolsky’s sign findings, along with the histopathologic report and other clinical findings that are consistent with SJS, the case published by Eloranta et al. [36] was still included in the analysis.
4.8 Additional Radiation-Related Considerations in SJS/TEN Cases
Several studies fell short of meeting inclusion criteria but are worth discussing. A case of a 67-year-old female with SJS/TEN overlap and a positive Nikolsky’s sign reported that sun exposure was a potentiating factor, providing images of the patient’s rash with photo-demarcated borders [10]. This case was excluded due to limited case information (only description was in a figure legend), and no causal drug or timeline of sun exposure was provided. Two studies reported patients with past histories of psoriasis who regularly received treatments of ultraviolet-B light and experienced fatal cases of TEN, one after seven days on cefozopran and the other 16 days after starting etretinate [131,132]. These cases did not meet inclusion criteria because the studies did not report whether the patient received ultraviolet-B light therapy after starting the causal drug. Three additional cases of SJS/TEN, which reported patients who received methoxsalen with concomitant UVR for the treatment of vitiligo or psoriasis [133], were excluded for similar reasons.
There are also reports of UVR exposure preceding introduction of the causal drug that resulted in SJS/TEN presentations. A patient who experienced a bullous sunburn on his back and thighs presented seven months later with a case of TEN after taking trimethoprim-sulfamethoxazole; his rash replicated the distribution of the previous sunburn [125]. Another study reported a man with a history of a left forearm sarcoma resection and treated with a split-thickness skin graft and radiotherapy who presented with SJS several months later [126]. His rash was localized primarily to the radiation exposure site following a 10-day regimen of trimethoprim-sulfamethoxazole for an infection. These were considered recall-like reactions as the areas of radiation exposure expressed significantly more vesiculobullous involvement than other areas of the body [126]. A recall reaction occurs after ionizing or ultraviolet radiation damages the skin, permanently destabilizing the skin’s immune behavior, leaving the affected skin compromised and vulnerable to subsequent immune-related disorders [134]. Recall reactions likely have a different pathogenesis and thus are considered a distinct diagnosis from photodistributed SJS/TEN cases that are described in this review.
Radiotherapy has also been reported to cause drug-induced SJS/TEN presentations after the initiation of a new drug regimen, similar to the UVR cases described in this review. One such case of SJS was reported after a patient received concomitant phenobarbital and radiation therapy, which resulted in rash distribution limited to the radiation exposure areas [135]. The same study recognizes an additional 21 cases of atypical erythema multiforme, TEN, and SJS in patients receiving radiation therapy while on concomitant phenytoin, phenobarbital, or carbamazepine. Additional studies described SJS/TEN overlap from concurrent gemcitabine or temozolomide and radiotherapy [136,137]. These cases were not included as all cases did not adequately describe rash onset, distribution, and type of radiation exposure. Future cases of SJS/TEN that are precipitated by radiotherapy should be evaluated by the proposed diagnostic criteria to determine if they are photodistributed SJS/TEN.
4.9 Limitations and Future Directions:
There were many studies that met inclusion criteria. However, one limitation to this review is the level of detail provided by the individual studies, thus limiting the analyses and comparisons that could be performed. Furthermore, it is difficult to ascertain the prevalence of photodistributed SJS/TEN. In the literature, UVR’s role in SJS/TEN cases is not widely documented. Due to the minimal coverage, it is plausible that many UVR-linked cases have gone unrecognized or unreported.
Future directions should focus on further investigation of the pathophysiology that is responsible for photodistributed SJS/TEN, as the current understanding is theoretical.