Keywords

1 Introduction

Lichenoid disorders are inflammatory dermatoses characterised clinically by flat-topped, pruritic, papular lesions and histologically by a band-like infiltrate of lymphocytes in the papillary dermis. Lichen planus (LP) is the most typical and well-recognised of the lichenoid dermatoses and presents with firm, shiny, polygonal, 1–3 mm papules with a red to violaceous colour and overlying fine white lines known as Wickham’s striae. Grey-brown pigmented macules may result upon resolution of primary lesions. On mucosal surfaces, Wickham’s striae are also often seen (Tziotzios et al. 2018; Shiohara and Mizukawa 2018).

LP-like or lichenoid drug eruptions (LDE) may be difficult to distinguish clinically and histologically from classic LP (Tziotzios et al. 2018; Shiohara and Mizukawa 2018; Ardern-Jones and Lee 2016). Identification of a drug cause may be difficult as the latency period between drug administration and onset of rash is variable and may be prolonged, up to several months or even years (Halevy and Shai 1993). Furthermore, resolution of the rash after discontinuation of the causative drug may take weeks to months (Halevy and Shai 1993), adding to the uncertainty of the diagnosis. Causality may be confirmed by re-exposure to the drug, but may not be acceptable to the patient.

2 Epidemiology

LDE is generally uncommon though specific reports on incidence rates are lacking. In fact, most epidemiological studies on cutaneous adverse drug reactions (CADRs) do not mention LDE. One study reported that LDE accounted for only 4% of all CADRs in a tertiary hospital in India (Qayoom et al. 2015). Approximately 10% of all LP cases are drug induced (Ardern-Jones and Lee 2016).

Age of presentation does not differ much between LDE and LP, reportedly ranging from 44 to 66 years for LDE (Halevy and Shai 1993; Lage et al. 2012; Fessa et al. 2012; West et al. 1990) and 47 to 50 years for LP (Lage et al. 2012). Paediatric LDE is rare (Payette et al. 2015) but may result from childhood vaccinations. There is reportedly no gender bias (Tziotzios et al. 2018).

3 Description of Features

Clinical and histological features which distinguish LDE from LP are summarised in Table 1. LDE tends to present with LP-like lesions (Fig. 1) which are more generalised, polymorphous, lack Wickham’s striae and have a more eczematous or psoriasiform appearance (Fig. 2). Photodistribution is more common in LDE and may be a useful diagnostic clue (Shiohara and Mizukawa 2018). Mucosal involvement (Fig. 3) is less common in LDE than in LP (Shiohara and Mizukawa 2018).

Table 1 Clinical and histological differences between lichen planus and lichenoid drug eruption
Full size table
Fig. 1
A photo of a hand with many bright spots. The dorsal side of the hand has more spots than the arms.

Lichen planus-like papules and plaques on the dorsal hand and forearm

Full size image
Fig. 2
A photo of an arm and elbow region. There are multiple bright spots all over the skin.

Lichenoid drug eruption with a more polymorphous appearance with mixture of LP-like and eczematous papules and plaques

Full size image
Fig. 3
A photo of an open mouth with visible red marks in the buccal cavity and plaque over the teeth.

White plaques with lace-like pattern and erosions on the buccal mucosa

Full size image

Differential diagnoses to consider include LP-like contact dermatitis [e.g. to methacrylic acid esters (Kawamura et al. 1996) and dimethyl fumarates (Guillet et al. 2009)], lichenoid keratosis (Pitney et al. 2016), paraneoplastic pemphigus (Tey and Tang 2009; Lim et al. 2018) drug-induced subacute cutaneous lupus (Crowson and Magro 1999), dermatomyositis (Al-Najjar et al. 1985), graft-versus-host disease (Hymes et al. 2006) and secondary syphilis (Tang et al. 2004).

4 Drug Causality

Arsenic was the first drug reported to cause LDE in 1929 (Almeyda and Levantine 1971). Since then, LDE has been reported to a long and growing list of drugs (Table 2). Many commonly used drugs, for example, beta-blockers, thiazide diuretics, angiotensin-converting enzyme inhibitors and non-steroidal anti-inflammatory drugs (NSAIDs), are now recognised to cause LDE. Drugs which have been recognised to be associated with LDE at certain anatomical locations, for example, sun-exposed areas or mucosa, are listed in Table 3.

Table 2 Drugs implicated in lichenoid drug eruptions
Full size table
Table 3 Drugs implicated in photodistributed and oral mucosal lichenoid drug eruption
Full size table

Older drugs such as gold (Penneys et al. 1974, Glenert 1984, Russell et al. 1997), penicillamine (Seehafer et al. 1981) and anti-malarials [e.g. quinine (Dawson 1986), quinacrine (Bauer 1981), chloroquine (Savage 1958)] were well-recognised to cause LDE but have become less commonly used.

Relatively new drugs which affect the immune system have now been reported to cause LDE. These include vaccines, for example, hepatitis B (Saywell et al. 1997; Rebora et al. 1999; Ferrando et al. 1998; Calista and Morri 2004; Schupp and Vente 1999; Limas and Limas 2002; Al-Khenaizan 2001) and human papillomavirus vaccines (Laschinger et al. 2015), interferon (IFN) therapy (Bush et al. 2017) and anti-HIV therapy [e.g. efavirenz (Baumrin et al. 2018), tenofovir (Gupta et al. 2015)]. In recent years, biological therapies [e.g. tumour necrosis factor [TNF] inhibitors (Inoue et al. 2017; El Habr et al. 2014; De Simone et al. 2008; Darrigade et al. 2016; Andrade et al. 2015; Utsu et al. 2012; Gonzalez et al. 2018)], targeted oncological drugs [e.g. imatinib (Sendagorta et al. 2009; Gómez Fernández et al. 2010; Sudha et al. 2011; Ena et al. 2004; Dalmau et al. 2006; Pascual et al. 2006; Lim and Muir 2002; Kawakami et al. 2009; Kuraishi et al. 2010)] and immune checkpoint inhibitors (Hwang et al. 2016; Cogen et al. 2018; Min Lee et al. 2018; Coleman et al. 2019; Curry et al. 2017; Shi et al. 2016; Obara et al. 2018; Biolo et al. 2019; Siegel et al. 2018; Coscarart et al. 2019) have featured as new and emerging causes of LDE.

4.1 Biologics

While TNF-inhibitors have well-known clinical efficacy in treating inflammatory conditions, it is now recognised that they may cause paradoxical inflammatory skin reactions (e.g. psoriasis). Numerous cases of LDE due to anti-TNFs have been reported (Inoue et al. 2017; El Habr et al. 2014; De Simone et al. 2008; Darrigade et al. 2016; Andrade et al. 2015; Utsu et al. 2012; Gonzalez et al. 2018). Interestingly, LDE has been reported to develop in a patient after switching from infliximab to its biosimilar, suggesting possibly different immunogenicity of the biosimilar drug (Gonzalez et al. 2018). Change in therapeutic class of biologics may still re-elicit the reaction, as reported in a psoriasis patient who developed LDE to an anti-TNF biosimilar with resolution after cessation but recurrence after introduction of an anti-IL17A drug (Maglie et al. 2018). LDE affecting oral mucosa due to anti-IL17A drugs have been reported (Thompson et al. 2016; Komori et al. 2017).

There are reports of LDE to anti-CD20 drugs which occurred in patients with follicular lymphoma (Kuten-Shorrer et al. 2014; Bakkour and Coulson 2012) and one of possible photodistributed LDE to rituximab in a patient with systemic lupus erythematosus (O'Connor et al. 2017).

4.2 Immune Checkpoint Inhibitors

Immune checkpoint inhibitors such as anti-cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) (e.g. ipilimumab), anti-programmed cell death 1 (PD-1) (e.g. nivolumab or pembrolizumab) and anti-programmed death ligand 1 (PD-L1) inhibitors (e.g. atezolizumab) are new therapies which activate the immune system against cancer cells and have demonstrated remarkable clinical efficacy. However, cutaneous toxicity is a common side effect and may be seen in up to 49% of pembrolizumab-treated (Hwang et al. 2016) and 60% of ipilimumab-treated (Min Lee et al. 2018) patients.

17–25% of all cutaneous toxicities in patients treated with anti-PD-1 drugs are lichenoid reactions (Hwang et al. 2016; Coleman et al. 2019). Interestingly anti-CTLA-4 drugs, and even anti-PD-L-1 drugs do not seem to cause lichenoid reactions as frequently as anti-PD1 drugs despite similar mechanism of action (Min Lee et al. 2018; Curry et al. 2017). Peripheral blood eosinophilia is only seen in 20% of anti-PD-1-induced lichenoid reaction. The mean time to onset is 88 days (range 1–266 days) and rash may even occur after discontinuation of the treatment (Hwang et al. 2016).

5 Variations in Clinical Features of LDE

Erosive LDE afflicting the oral or genital mucosa is not uncommon. Drugs implicated include beta-blockers (Fessa et al. 2012), anti-PD-1 drugs (Shi et al. 2016; Obara et al. 2018), lithium (Srebrnik et al. 1991; Hogan et al. 1985), NSAIDs (Hamburger and Potts 1983) and sulphonylureas [e.g. glimepiride (Hammami et al. 2015)].

Cutaneous blisters are rarely associated with LDE and have been reported in cases of LDE to naproxen (Güneş et al. 2006), labetolol (Gange and Jones 1978), radiocontrast media (Grunwald et al. 1985) and tiopronin (Hsiao et al. 1986). Anti-PD-1 drugs may occasionally cause blisters (Biolo et al. 2019) but clinicians should also consider differential diagnoses of bullous pemphigoid or Stevens–Johnson Syndrome in such cases (Siegel et al. 2018).

Hypertrophic (Coscarart et al. 2019) and linear (Utsu et al. 2012; Gencoglan et al. 2009) forms of LDE have been rarely reported.

Nail changes are rarely reported in LDE but are similar to those in LP and include longitudinal ridging, onychoschizia and dorsal pterygium (May et al. 2017; Zheng et al. 2017). Subungual hyperkeratosis has been reported in LDE to imatinib (Dalmau et al. 2006). Interestingly, one patient developed LDE to propylthiouracil with only nail changes (red nodules on the nail bed) without lesions on skin or mucous membrane (Saito et al. 2007).

Scarring alopecia has been reported in a patient with lichen planopilaris with concurrent oral erosive LDE induced by pembrolizumab (Cogen et al. 2018).

Other rare associations with LDE include decreased sweat production with atrophic sweat glands in quinacrine-induced LDE (Sulzberger et al. 1947) and palmoplantar hyperkeratosis in imatinib-induced LDE (Kuraishi et al. 2010).

6 Histological Findings

The most characteristic histological feature of both LDE and LP is lichenoid interface dermatitis which is a band-like lymphocyte infiltration of the papillary dermis associated with apoptosis of the basal keratinocytes.

The “classical” histopathologic findings that are indicative of LDE are eosinophils and plasma cells in the cellular infiltrate, focal parakeratosis, and an infiltrate around deep vessels (Figs. 4 and 5) (Van den Haute et al. 1989). Lage et al. reported that focal parakeratosis, focal interruption of the granular layer and cytoid bodies (representing apoptotic keratinocytes) in the cornified and granular layers were present in more than 50% of LDE and never in idiopathic LP (Lage et al. 2012).

Fig. 4
A histological study of the cells. It has a dark upper margin and lightly stained cytoplasm. The skin surface is wedge-shaped and cornified.

Histology of lichenoid drug eruption. The epidermis is acanthotic with wedge-shaped hypergranulosis, “saw-toothed” rete ridges and focal parakeratosis. Civatte bodies are present within the epidermis. There is a dense band-like lymphocytic infiltrate in the upper dermis associated with vacuolar alteration of the basal keratinocytes (haematoxylin and eosin, low magnification)

Full size image
Fig. 5
A histological study of cells with prominent dark dyed cells all over the surface. The cells have spaces between them.

Inflammatory infiltrate of lichen drug eruption consisting of lymphocytes and numerous eosinophils (haematoxylin and eosin, high magnification)

Full size image

Histological findings in photodistributed LDE may be indistinguishable from those of idiopathic LP and that a biopsy specimen which shows the classic features of LP should not be used as evidence against a drug eruption, especially if the lesions are photodistributed (West et al. 1990).

Histologic features of anti-PD-1-induced lichenoid reaction have been reported to be polymorphous, that is, one lesion may have features of LDE while other lesions may demonstrate other histological patterns (e.g. spongiotic dermatitis) (Tetzlaff et al. 2017).

Immunohistochemistry demonstrates that the inflammatory infiltrate is predominantly of CD8 cytotoxic cells. The number of granzyme B-expressing cells is reported to be positively correlated with degree of keratinocyte apoptosis (Lage et al. 2012).

Giant cell lichenoid dermatitis is an uncommon histologic variant first reported in 1986 by Gonzalez et al. (1986). It is characterised by granulomatous inflammation in the dermis in addition to the usual features of LP. Groups of histiocytes, with or without giant cell formation, are seen in the mid to reticular dermis or admixed with the lichenoid inflammatory cells. This particular histological pattern is not associated with specific drugs. A wide range of drugs have been implicated, for example, beta-blockers, calcium channel blockers, angiotensin-converting enzyme inhibitors, statins, proton pump inhibitors, nonsteroidal anti-inflammatory drugs and sulphonamide antibiotics (Magro and Crowson 2000; Braswell et al. 2019).

There are no significant differences in direct immunofluorescence (DIF) staining between LDE and idiopathic LP (Halevy and Shai 1993). DIF reveals “shaggy” band-like deposits of fibrinogen along the dermoepidermal junction, well as colloid bodies staining with any of the autoantibodies immunoglobulin M (IgM), IgG, IgA and C3 with or without fibrin. DIF remains useful to distinguish LDE from auto-immune conditions.

7 Pathogenesis

Mechanisms leading to the development of LDE have not been well elucidated. It is likely that the condition occurs in a predisposed individual when the causative drug triggers off immune dysregulation in a conducive micro-environment.

While genetic factors such as human leucocyte antigen (HLA) haplotype have been strongly associated with other CADR, these have not been well-investigated in LDE. Studies performed in oral LDE have also failed to show a pathogenic role for polymorphisms in cytochrome P450 enzymes which may influence drug metabolism (Kragelund et al. 2009, 2010).

It has been proposed that peripheral blood lymphocytes recruited during an immune or inflammatory reaction (e.g. virus infections) could remain in the skin as resident memory T cells. These memory T cells could be reactivated, cross-reacting with different agents, resulting in localised damage of the epithelium (Giuliani et al. 2008).

Ultraviolet radiation (UVR) may play a contributory role as seen by the photo-distribution commonly seen in LDE and involvement of drugs such as anti-malarials, thiazides and statins, which are well-recognised photosensitisers. UVR may induce free radical production by drugs, resulting in cellular damage leading to an inflammatory cascade involving various cytokines and inflammatory cells (Khandpur et al. 2017).

The role of type 1 interferon in pathogenesis of LDE has been suggested by the occurrence of LDE in patients treated with interferon-α and TNF-α inhibitors. It has been proposed that TNF-α inhibition allows for upregulation of IFN-α and in turn, IFN-α induces activation of resident T cells and plasmacytoid dendritic cells, mediates recruitment of cytotoxic T cells and upregulates the expression of cytotoxic agents (e.g. perforin) by cytotoxic T cells and NK cells (Asarch et al. 2009).

LDEs induced by beta-blockers may offer another clue to pathogenesis. Beta-adrenoreceptors are found in keratinocytes, Langerhans and dendritic cells and have a role in controlling the Th1 response to pathogens. Beta-adrenergic dysfunction has been reported in keratinocytes in psoriasis and vitiligo lesions (Sivamani et al. 2007). Likewise, beta-blockers may theoretically result in sustained inflammatory reaction.

The PD-1/PD-L1 pathway plays a vital role in inhibitory control of T lymphocytes. PD-1 inhibitors may cause lichenoid reactions by unleashing the immune response to an antigen in the skin or alternatively, by unmasking an immune response to another drug which was previously tolerated (Shi et al. 2016).

7.1 Treatment

Identification of causative agent may be difficult in patients taking multiple chronic medications. Dietary exposure must not be neglected as quinine in tonic water (Russell et al. 1997) and gold in certain alcoholic beverages (Russell et al. 1997) have been reported to cause LDE. Stopping the causative drug typically results in resolution of the lesions over a period of weeks to months. Patch tests are of low sensitivity in LDE (Osawa et al. 1990) but may be considered if there is uncertainty about which drug to stop. There have been reports of LDE which improved or resolved completely despite continuation of the causative drug (Asarch et al. 2009).

Treatment with topical steroids is usually beneficial with occasional cases requiring systemic steroids. Acitretin has been reported to be useful in treating LDE due to imatinib, allowing continuation of treatment in a cancer patient (Asarch et al. 2009).

For lichenoid reactions induced by anti-PD-1 and anti-PD-L-1 drugs, the condition is usually not severe and with appropriate management, only a small percentage (< 10%) require interruption of treatment (Coleman et al. 2019; Shi et al. 2016). Continuation of treatment is generally favoured as the occurrence of immune-related adverse events and dermatologic reactions appears to be associated with more favourable oncologic outcomes (Min Lee et al. 2018; Sanlorenzo et al. 2015; Chan et al. 2019). Nevertheless, clinicians should remain aware of potential complications of oral mucosal LDE. Just as classic oral lichen planus has potential for malignant transformation, squamous cell carcinoma has been reported in a case of mucosal LDE to pembrolizumab (Owosho et al. 2016).

In conclusion, diagnosis and management of lichenoid drug eruptions is a challenging task for the clinician. Keeping up to date with developments in new drugs remains crucial.