Epidermal growth factor receptor inhibitors (EGFR/HER1/Erb1) are antitumor agents that have emerged in recent years to treat advanced-stage solid tumors.1,2 Inhibition of EGFR-mediated signals is achieved with 1) monoclonal antibodies, namely cetuximab (Erbitux) and panitumumab (Vectibix), which are used mainly in metastatic colorectal cancer,1,3 but also in squamous cell carcinoma of the head and neck and skin2,4 and 2) low-molecular-weight drugs, namely erlotinib (Tarceva), gefitinib (Iressa), used in lung cancer and pancreatic cancer, and lapatinib (Tykerb), used in breast cancer.5 These drugs prevent adenosine triphosphate binding to the intracellular portion of EGFR. Unlike conventional chemotherapy, EGFR inhibitors are used at an optimal biological dose to achieve better tolerance.

EGFR is a transmembrane glycoprotein expressed in the skin and in the epithelial cells of 30% to 100% of solid tumors.2 The activation of EGFR by its growth factor ligands triggers a series of processes that regulate epidermal proliferation, differentiation, apoptosis, migration and synthesis of inflammatory cytokines.6 EGFR therefore plays a role in the development and normal differentiation of epidermal keratinocytes, stimulating their growth, protecting against UV-induced damage, inhibiting inflammation, and favoring wound healing. By altering keratinocyte proliferation and differentiation, EGFR inhibition exerts an antineoplastic effect, but it also triggers a series of cutaneous toxic effects, such as abnormal follicular keratinization and a secondary inflammatory response.5,6 All EGFR inhibitors cause dose-dependent cutaneous toxicity.7

Adverse cutaneous effects are observed in sites expressing high levels of EGFR, and the most common effect is a papulopustular rash, which occurs in over 80% of patients.7,8 The rash tends to be mild, but it can be moderate or severe in up to 50% to 60% of cases,9,10 requiring dose reductions or discontinuation of antitumor treatment. Other adverse effects are paronychia, hair alterations, pruritus, and xerosis.

Research is underway to study markers that help to predict the effectiveness of EGFR inhibitors and avoid unnecessary toxicity by limiting administration to patients who will potentially benefit from them. Activating KRAS and NRAS mutations are currently the only predictors of resistance to approved anti-EGFR inhibitors used in clinical practice, and their evaluation has been included in the summary of product characteristics of cetuximab and panitumumab by the European Medicines Agency.1

Other potential clinical and genetic markers of tumor response, however, have also been investigated. Genetic studies have shown that EGFR polymorphisms involved in gene regulation are associated with rash and tumor response, although no consensus has yet been reached on their predictive value.1,3,8,11 Our group recently suggested that the single nucleotide polymorphism-216 might be a useful predictor of tumor response in metastatic colorectal cancer.12 One of the main clinical markers of response to anti-EGFR monoclonal antibodies is the development of a papulopustular skin rash.3

The existence of clinical predictors for this skin rash, however, is less clear, although young age and male sex have been proposed as the main predictors.11,13

We performed a retrospective observational study of 116 patients with metastatic colorectal cancer under treatment with intravenous anti-EGFR monoclonal antibodies (cetuximab every week or panitumumab every 2 weeks) at Hospital Universitario Donostia between January 2010 and January 2013.

The following outcome variables were analyzed: papulopustular rash, face eczema, xerosis, pruritus, paronychia, fissures, hypertrichosis, trichomegaly, sores, herpes, and inflammation of sun-damaged skin. Skin rash severity was measured using the National Cancer Institute's modified Common Terminology Criteria for Adverse Events,2,14 which is the only standardized scale currently available. The independent variables analyzed were tumor response, age, sex, place of origin, type of anti-EGFR monoclonal antibody, associated chemotherapy, culture testing, discontinuation of antitumor treatment, and treatment with tetracyclines or isotretinoin. Tumor response was assessed using the Response Evaluation Criteria in Solid Tumors.12,14

Four patients were excluded from the analysis: 3 because they died of cancer before being evaluated by the dermatology department and 1 because cetuximab treatment was discontinued following an infusion reaction. Tumor response was classified as inevaluable in 13 patients and was therefore analyzed in 103 of the 116 patients studied.

We studied 116 patients (35 women and 81 men) aged between 31 and 86 years (mean, 64.9 years). The main epidemiological and clinical characteristics are shown in Table 1.

Ninety-five patients (81.9%) developed a papulopustular rash, which was classified as moderate to severe (grades 2-3) in 70.5% of cases.

The median time between initiation of anti-EGFR therapy and the appearance of the rash was 10 days (interquartile range, 7-15 days). No significant differences were observed for time of rash onset between cetuximab and panitumumab (median, 8 vs 10 days).

Culture was performed for 22 samples. Just 2 of these were positive for Staphylococcus aureus, supporting previous reports that these pustules are sterile. Skin biopsy was performed in 2 patients and revealed neutrophilic folliculitis in both cases.

Koebner phenomenon was observed in hospitalized patients who developed the rash on pressure areas and in patients with a reservoir or an ostomy.

In the univariate analysis, papulopustular rash (presence vs absence) was a predictor of tumor response, with incidence increasing with better tumor response. All the patients who achieved complete tumor response developed a rash, and the presence of the rash became less common with poorer tumor response (P=.03, Fig. 1). Although the association lost strength when the rash was classified by severity, it maintained a clearly linear trend (P=.05, Table 2).

Figure 1.

Association between tumor response and development of papulopustular rash. The incidence of rash was greater in patients who responded to anti-epidermal growth factor receptor (EGFR) therapy. All the patients who achieved complete tumor response developed the rash, but this became less common with worsening tumor response. Just 76.7% of patients with progressive disease developed a papulopustular rash (P=.03, χ2 test). RECIST indicates Response Evaluation Criteria in Solid Tumors.

(0,1MB).

Multivariate ordinal logistic regression analysis was used to investigate the capacity of the clinical variables analyzed to predict the development of a papulopustular rash (Table 2). In the univariate analysis, only number of treatment cycles and tumor response were significantly associated with the development of a skin rash. Number of cycles retained its significance in the multivariate analysis, with a higher number of cycles associated with a greater risk of more severe rash (P=.02). Severe rash was also associated with complete tumor response (P=.08), male sex (P=.07), and younger age (P=.09).

The papulopustular rash was treated as soon as possible to avoid having to discontinue anti-EGFR therapy due to toxic effects. Fig. 2 shows the treatment algorithm used at our department to treat papulopustular rash according to severity. All patients with a grade 2 or 3 rash were treated with oral tetracyclines (Fig. 3A), which achieved adequate control (Fig. 3B). Isotretinoin was needed to treat 3 patients with a more refractory skin rash, and a good response was obtained in 2 of these. The third patient did not tolerate this treatment, leading to discontinuation of anti-EGFR therapy.

Figure 2.

Treatment algorithm for papulopustular rash. Summary of treatment of EGFR inhibitor–induced skin rash used at our department based on reports in the literature and our clinical experience. Treatment is individualized in each case.

(0,32MB).

Figure 3.

Papulopustular rash before (A) and after (B) treatment with tetracyclines for 2 months. The rash improved after 2 months’ treatment with doxycycline 100mg/d. The rash remained similar in size but improved in terms of severity.

(0,19MB).

In the current series, neither oral tetracyclines nor isotretinoin altered tumor response.

The epidemiological characteristics of cutaneous toxicities other than papulopustular rash are shown in Table 3 and Fig. 4. We analyzed the relationship between these adverse cutaneous effects and sex, age, EGFR inhibitor, associated chemotherapy, and tumor response.

Figure 4.

Timeline and frequency of appearance of cutaneous toxicities. The papulopustular rash was the first cutaneous adverse effect to appear, and affected 81.9% of patients. The remaining effects started to appear after a month's treatment, but were less common (11%-40%).

(0,25MB).

Facial eczema was more common in patients treated with panitumumab than with cetuximab (23 vs 11, P=.02) and in some cases it was so severe it needed treatment with systemic corticosteroids.

Xerosis was more common in patients who responded to antitumor therapy and in elderly patients, although the difference did not reach statistical significance in the second group of patients. Thirty-nine (53.4%) of the 73 patients who achieved tumor response developed xerosis, compared with just 6 (20.0%) of those who did not (P=.002). The xerosis had an ichthyosis-like appearance in 6 cases. Pathologic examination was performed in 1 of the patients and showed lesions consistent with acquired ichthyosis.

No significant associations were found for any of the other cutaneous toxicities.

Cultures were ordered for 5 of the 21 paronychias to rule out secondary infection. Two were positive for S aureus, 1 was positive for Streptococcus pyogenes, and 1 was positive for Candida albicans.

The most common hair alterations were trichomegaly, hypertrichosis, and hirsutism (Table 3), although there were also 5 cases of diffuse hair loss (1 woman and 4 men).

Mucosal involvement was observed in 16 patients. Eleven developed herpes lesions on the oral and/or genital mucosa, but 10 of these were also receiving chemotherapy. Five of the patients undergoing chemotherapy had oral sores, and 2 of these needed treatment with systemic corticosteroids to control the sores.

Photosensitivity occurred in patients receiving chemotherapy with 5-fluorouracil derivatives. However, 3 patients being treated with panitumumab monotherapy experienced inflammation of existing sun-damaged skin. This association has been described very rarely with EGRF inhibitors.

It was possible to discontinue methotrexate therapy in 1 patient with psoriasis and psoriatic arthritis whose psoriasis improved during anti-EGFR therapy.

Finally, anti-EGFR therapy had to be discontinued in 8 patients (6.8%) due to cutaneous toxicity and to be delayed or reduced in an additional 21 patients (18.1%) (Table 4). The main reasons were papulopustular rash, fissures, and paronychia.

Although papulopustular rash can be considered an in vivo marker of anti-EGFR activity, the relationship between other EGFR inhibitor–induced cutaneous toxicities and tumor response has not been described in the literature.3,8 In the current series, we observed an association between xerosis and anti-EGFR therapy, with higher rates seen in patients who responded better to treatment. This is the first time this association has been described and we believe that, like papulopustular rash, xerosis may be a clinical marker of tumor response.

Our findings also confirm the previously described association between papulopustular rash and tumor response,3,12 with patients who responded better to anti-EGFR therapy showing an increased risk of severe rash. These results add strength to the idea that tumor response and cutaneous toxicity have a similar mechanism of action. Patients who develop more severe skin rash would appear to benefit more from anti-EGFR therapy. Several authors have suggested employing a dose escalation to rash strategy in patients who do not develop the papulopustular rash. This strategy involves administering increased doses of the drug until the patient develops the rash.15 Other authors, in turn, have suggested that the absence of the rash after a given time should prompt the discontinuation of treatment, but this is not a decisive factor in current clinical practice.1 In our opinion, and considering that a small subgroup of patients who do not develop a rash might actually benefit from anti-EGFR therapy, this decision should be based on a series of clinical and genetic markers that would permit more individualized therapy.

On investigating the association between the development of a papulopustular rash and other clinical predictors, we found a trend towards significance for age and sex, supporting findings by Jatoi et al.11 These authors reported that the rash was more common in male patients and younger patients, and hypothesized that older patients have lower levels of EGFR and that both androgens and estrogens interact with this receptor.

Numerous treatment options have been described over the years for cutaneous toxicity induced by EGFR inhibitors, but no consensus has yet been reached. Given the current lack of evidence-based guidelines, current treatment strategies are based on expert opinions10 and clinical experience.

Several topical treatments have been used, but without proven success,9,16–18 and just 3 randomized trials have been conducted. Scope et al.19 evaluated the use of topical pimecrolimus applied to half of the face, but found no clinical benefit, like Jatoi et al.,20 who evaluated the use of preventive treatment with a sunscreen for a month compared with placebo. Finally, Pinta et al.9 found that the application of vitamin K1 cream appeared to have a beneficial effect, but their comparative data were from reports in the literature (Table 5).

Considering other papulopustular conditions, such as rosacea and acne, oral tetracyclines, which have both anti-inflammatory and antiangiogenic properties—described by Fernandez-Guarino et al. in 20062,21—would be the first-line treatment for papulopustular rash. Favorable outcomes have been reported for azithromycin,22 isotretinoin,23 and acitretin.24 Isotretinoin is recommended for more refractory cases, as it is not clear how this drug might affect the mechanism of action of EGFR inhibitors. Accordingly, some authors have recommended limiting its use in this setting.7Table 5 summarizes the findings of 5 randomized clinical trials that have analyzed the prophylactic use of oral tetracyclines in papulopustular rash.14,25–28

Lacouture et al.14 compared the prophylactic and active use of doxycycline to improve panitumumab-induced cutaneous toxicity and found a 50% reduction in the incidence of moderate to severe rash (Table 5).

Jatoi et al., Scope et al., and Deplanque et al. compared, respectively, the use of oral tetracyclines, minocycline, and doxycycline with placebo from the start of anti-EGFR therapy, and found that these drugs decreased the severity and number of lesions, but not the incidence of the rash.26–28 In a more recent report, Jatoi et al. concluded that tetracyclines were not an effective prophylactic treatment for papulopustular rash,26,28 and Scope et al.27reported no additional benefits after 8 weeks of treatment (Table 5). Given the controversy surrounding the effectiveness of tetracyclines, which was highlighted in a recent meta-analysis,28 these drugs cannot be recommended as preventive standard therapy in EGRF inhibitor–induced rash. Nevertheless, according to the American clinical practice guidelines for the prevention and treatment of EGFR inhibitor–associated dermatologic toxicities doxycycline 100mg should be administered every 12 to 24hours for the first 6 weeks of treatment.10 The treatment algorithm used in our hospital establishes tetracyclines as a first-line treatment for papulopustular rash, as in our experience they achieve good control. We usually maintain the treatment for the duration of anti-EGFR therapy, or for less time if the rash is controlled sooner. No specific studies have analyzed treatments for other cutaneous toxicities induced by EGFR inhibitors, although doxycycline has been proposed as a potential option for paronychia. Management strategies typically include hygiene measures, emollients, and topical antiseptics, antibiotics, and corticosteroids.10 The use of sun protection measures is also important, as some patients experience inflammation of existing sun damage, possibly due to the high levels of EGFR that tend to be found in the follicular ostium and in precursor skin cancer lesions.29

Finally, although there have been reports of worsening psoriasis,30 1 patient in our series experienced improvement of his psoriasis during anti-EGFR therapy, possibly due to the antiproliferative effect of EGFR inhibition, as hyperproliferation has been implicated in the pathogenesis of this skin condition.7

The results of our study suggest that, like papulopustular rash, xerosis may be a predictor of tumor response in patients treated with EGFR inhibitors. We therefore believe that both markers should be used to identify patients who might benefit from anti-EGFR therapy.

Furthermore, treatment of papulopustular rash should be initiated early and continued, as suboptimal treatment can lead to treatment delays or interruption in patients who are benefiting from anti-EGFR therapy.

In our experience, treatment should be individualized and adjusted over time according to the course of the different cutaneous adverse effects.

The authors declare that they have no conflicts of interest.