Topically administered tacrolimus does not increase cancer risk in children with atopic dermatitis

Transdermal absorption of tacrolimus

In the pharmacoepidemiological study APPLES (A Prospective Pediatric Longitudinal Evaluation to assess the long-term Safety of tacrolimus ointment for the treatment of atopic dermatitis), which responded to this challenge from regulatory authorities, the effect of tacrolimus was studied. This calcineurin inhibitor is minimally absorbed when applied dermally. According to pharmacokinetic studies, plasma concentrations after its administration reach < 1.0 ng/ml, while achieving 5–20 ng/ml is required to induce systemic immunosuppression. Therefore, the systemic effect is likely minimal, but the possible adverse impact of tacrolimus on skin immune cells and processes remains a question.

Pharmacoepidemiological study

The prospective observational cohort study APPLES, conducted in 9 countries and lasting 10 years, focused on quantifying malignancies in children with AD treated with tacrolimus for more than 6 weeks. The standardized incidence rate in these children was then compared to the usual incidence rate in a comparable population in terms of age, gender, and race of the children. These data were obtained from national cancer registries.

Incidence of malignancies in the studied population

The study included 7,954 patients. During 44,629 patient-years, 6 cases of malignant cancer were recorded and confirmed. The expected number of cases calculated from national registries data was 5.95. Identified were spitzoid melanoma, alveolar rhabdomyosarcoma, appendiceal carcinoid, spinal tumor, malignant paraganglioma, and chronic myeloid leukemia. No cases of lymphoma or non-melanoma skin cancer were reported. The period from starting tacrolimus treatment to cancer diagnosis ranged from 4.1 to 10.9 years. The standardized incidence rate was 1.01 (95% confidence interval 0.37–2.20), thus no increased risk was confirmed.

Statistical sensitivity analysis and discussion of the results

Childhood cancers are rare, requiring a large number of study participants to record enough events for statistical analysis. High patient loss during long-term follow-up is a typical limitation of longitudinal observational studies. The APPLES study also experienced losses. A total of 4,368 patients, accounting for 54.9% of the studied population, stopped communicating with the trial physicians. Additionally, 7 children died (due to injury or severe systemic disease unrelated to treatment) and 1,454 participants withdrew consent to participate in the study.

However, such patient loss is accounted for when planning long-term studies. The authors conducted a statistical sensitivity analysis, which can help partially fill the gap left by the data from 'lost' patients. Theoretically, it is possible that the 'lost' children had a higher incidence of cancer, and the loss of this data could influence the overall study outcome. The analysis showed that there would need to be 2.7 times more malignant tumors in these patients for the entire patient cohort's result to exceed the expected incidence calculated from cancer registries. Therefore, a significant distortion due to the loss of followed patients probably did not occur.

Conclusion

The incidence of malignant diseases was comparable to the incidence in the pediatric population not treated with tacrolimus. The identified malignant disease cases were diverse, had different etiologies, and did not seemingly relate to the duration of tacrolimus administration. The prospective longitudinal APPLES study found no evidence of increased cancer risk in children with AD treated with topically administered tacrolimus.

(jam)

Sources:
1. Paller A. S., Fölster-Holst R., Chen S. C. et al. No evidence of increased cancer incidence in children using topical tacrolimus for atopic dermatitis. J Am Acad Dermatol 2020; 83 (2): 375–381, doi: 10.1016/j.jaad.2020.03.075.
2. SPC Protopic. Available at: www.ema.europa.eu/en/documents/product-information/protopic-epar-product-information_cs.pdf