Voriconazole-induced photosensitivity most commonly presents as a “sunburn” response on sun-exposed surfaces of the body; it may also present as cheilitis, exfoliative dermatitis, pseudoporphyria cutanea tarda, or lesions resembling discoid lupus erythematosus.5,6,8,12
In general, these reactions are reversible upon discontinuation of the drug. Recent reports have also described pigmentary changes limited to the photo-exposed surfaces of the skin in voriconazole-treated pediatric patients, suggestive of accelerated photoaging and chronic photodamage.2,13
Of greatest concern are three single international case reports of aggressive and multifocal cutaneous squamous cell carcinoma with voriconazole use in the immunocompromised setting (renal allograft recipient, HIV infection, and chronic granulomatous disease.)9–11
In addition, preliminary abstract data suggests that lung transplant patients taking voriconazole antifungal prophylaxis may be at increased risk of non-melanoma skin cancer.14
Chronic immunosuppression is a well-recognized risk factor for the development of non-melanoma skin cancer. SCC in particular is a significant long-term complication for many renal allograft recipients.15
However, the short duration of immunosuppression preceding the development of SCC, the unusually high number of SCC tumors following SCT, and the young age at onset of SCC observed in this series suggest that voriconazole-associated phototoxicity accelerates the risk of SCC formation in the immunocompromised setting.
The mechanism of voriconazole-induced photosensitivity is not clearly understood, but may result from a metabolite of the drug, rather than the drug itself. Although voriconazole does not absorb in the UVA or UVB spectrum, its major metabolite, voriconazole N-oxide, absorbs in the UVB and UVA ranges,16
and may therefore act as the requisite chromophore for phototoxicity. Voriconazole is extensively metabolized by the cytochrome P450 enzymes CYP2C19, CYP2C9, and CYP3A4, and less than 2% of the drug is excreted unchanged.17
Polymorphisms in CYP2C19 significantly influence voriconazole metabolism.17
Homozygous poor metabolism polymorphisms in CYP2C19 result in plasma voriconazole concentrations 3–5 times that found in extensive metabolizers. The prevalence of homozygous poor metabolizer CYP2C19 polymorphisms in the population is 20–30% in Asians and 2–3% in Caucasians.17
To our knowledge, a relationship between voriconazole metabolism and phototoxicity has not been reported, and further investigation is needed to determine the variability in photosensitivity observed between individual patients.
All of the patients in this series demonstrated one or more cutaneous signs of phototoxicity, implicating ultraviolet radiation-induced DNA damage in promoting cutaneous malignancy. In addition to direct UVA-induced DNA damage, there is evidence that UV-induced alteration of the microenvironment (cell-cell interactions, cytokine release, cell-extracellular matrix interactions, inflammation, and loss of T regulatory cells) may also promote the development of cutaneous malignancy.18
The extensive lentigo formation seen in this series is reminiscent of a similar phenomenon noted in patients undergoing phototherapy with psoralen (a photosensitizer) and ultraviolet A (PUVA). Patients receiving high cumulative doses of PUVA are at greatly increased risk of SCC formation, an effect that persists even after discontinuation of the treatment.19
There is also experimental evidence that phototoxic drug exposure is associated with an elevated risk of cutaneous malignancy. Photocarcinogenesis studies utilizing fluoroquinolones have demonstrated accelerated skin tumor formation in mice, even when sub-phototoxic doses of UVA radiation were administered chronically.20
A correlation between photosensitizing drug exposure and non-melanoma skin cancer was also identified in two recent population-based case-control studies.21,22
Although we cannot completely discount the potential contribution of other potential photosensitizing agents (TMP/SMX, dapsone) administered concurrently with voriconazole in four patients in this series, it should be noted that other signs of premature photoaging (lentiginoses, actinic keratoses) and SCC are not typically encountered with the use of these agents.
Although prolonged immunosuppressive therapy and severity of GVHD are known risk factors for the development of SCC in the HCT setting,23
in our 5-year experience systematically evaluating approximately 150 treatment-refractory patients with chronic GVHD at the NIH, cutaneous SCC is uncommon, despite chronic immunosuppression employed in nearly all patients. In contrast to the organ transplant setting, the cumulative incidence of oral and cutaneous SCC in a cohort of 24,011 HCT recipients was only 1.1% at 20 years (95% CI: 0.7–1.7).23
Similarly, a 20-year follow-up of 4,810 allogeneic HCT recipients at the Fred Hutchinson Cancer Research Center yielded a cumulative SCC incidence estimate of only 3.4%.24
Both the early age of onset (median 34.5 years; range 9–54) and short duration of immunosuppression (median 51 months; range 13–122) prior to the development of skin cancer in our series are also highly unusual in comparison to established data on SCC incidence following immunosuppression in the renal allograft and HCT settings. The latency between renal transplantation and first skin cancer formation is approximately three years in allograft recipients over age 60, but increases to 8 years in patients younger than 40 years.15
In the Curtis23
bone marrow registry review of 24,011 HCT patients, the median interval between HCT and SCC diagnosis was 7 years. Similarly, at Fred Hutchinson Cancer Research Center, SCC developed a median 6.3 years after HCT in a somewhat older population (median age 48.9 years).24
In comparison, the median duration to first SCC amongst the 6 HCT recipients in this series was only 4.1 years, despite a younger age (median 34.5 years, range 9–54) at time of transplantation.
Even with the apparent link between photosensitizing drugs and cancer risk suggested by both this study and preliminary population based-data, a complete understanding of the role of phototoxic drugs in SCC development in individual patients, particularly those who are immunosuppressed, is unclear. Most phototoxic drugs, including TMP/SMX and flouroquinolones, are typically prescribed for a limited time period, and alternative regimens are readily available if severe photosensitivity is detected. By contrast, voriconazole is often employed as chronic therapy for ambulatory, immunocompromised patients for whom oral alternatives with similar broad-spectrum anti-fungal activity were not available until recently. Thus, SCC development may require chronic phototoxicity as well as a “high-risk” immunocompromised population in a manner analogous to the increased risk of SCC following long-term PUVA phototherapy in patients who also receive cyclosporine.25
We believe a high index of suspicion for photosensitivity is warranted when using voriconazole, particularly in light of the potential for misdiagnosis of this reaction as GVHD in the post-HCT setting.2
Until the role of voriconazole in the development of SCC in the immunocompromised setting is more clearly understood, we recommend strict photoprotective measures be employed. Careful reappraisal of the need for long-term voriconazole prophylaxis may be of value, particularly in patients who manifest signs of chronic photodamage or have a history of skin cancer.