Testimony of Sidney M. Wolfe, M.D.
Director, Health Research Group of Public Citizen
Before the FDA Pediatric Advisory Committee:
Additional Evidence of the Cancer Risk of Calcineurin Inhibitor Skin Medications
Tacrolimus (Protopic) and Pimecrolimus (Elidel)
- I will first briefly review the newly published results of the largest epidemiological study to examine the carcinogenicity of these drugs, a study not included in the briefing materials, probably because it was published after the briefing materials were assembled.
- Background information about animal carcinogenicity of these drugs and earlier FDA concerns will be discussed.
- I will then summarize the additional findings from FDA’s newer cases of pediatric malignancies reported to the AERS system and place them in the context of the growing evidence of the T-cell lymphoma carcinogenicity of these drugs.
- Finally, I will recommend much more stringent regulatory action than has been proposed by the FDA. The risk benefit ratio for treating eczema or atopic dermatitis with these dermal preparations will be contrasted with the risk benefit ratio for immune suppressive drugs such as tacrolimus, used orally or by injection to inhibit organ rejection.
1. New Epidemiologic Evidence
In December, a cohort study based on 953,064 subjects with diagnoses of atopic dermatitis or eczema between 2001 and December 2004 in the Kaiser Permanente California data base, was published. This was the largest cohort study conducted to evaluate the association between calcineurin inhibitors and cancer. The goal of this cohort study was to compare the risk rates of overall cancer and of different types of cancer among patients with eczema or atopic dermatitis who were exposed to topical tacrolimus or pimecrolimus versus those who were unexposed. 914,382 (96%) did not receive topical tacrolimus or pimecrolimus during the study period. The remaining 4% of the cohort (n = 38,682) were exposed to topical tacrolimus alone (n = 11,898), pimecrolimus alone (n =22,716), or both (n = 4068). Those exposed to either drug included 17,807 who were 20 years or younger or 42% of the total 42,750 people who were exposed. (6,663 or 16% were 0 to two years of age). No specific age information was provided for the cancer cases.
Although there was no significant increased risk of overall cancers, the study found a statistically significant increased risk in T-cell lymphoma. There were 16 unique cases of T-cell lymphoma; 8 cases were exposed to tacrolimus alone, 6 were exposed to pimecrolimus alone, and 2 were exposed to both drugs prior to the cancer diagnosis. Among the 16 cases, 13 (81%) were Cutaneous T–Cell Lymphoma (CTCL), and the other 3 cases were located at lymph nodes, lungs, and small intestines.
A chart review that revealed 4 of the 16 cases were suspected of having pre-existing CTCL. For the remaining 12 cases of T Cell Lymphoma, 7 patients were exposed to tacrolimus only, 3 were exposed to pimecrolimus only, and 2 were exposed to both drugs. After these 4 cases were excluded, the age and sex hazard ratio for T-cell lymphoma was 5.44 (95% CI 2.51 to 11.79; p < 0.001) for tacrolimus and 2.32 (95% CI 0.89 to 6.07; p = 0.086) for pimecrolimus. The majority (89%) of patients taking tacrolimus were exposed to the higher strength ointment (0.1%), while only 2 (22%) cases were exposed to the lower strength (0.03%). The authors pointed out that the statistically significant increase with tacrolimus but not quite significant increase with pimecrolimus are consistent with the fact that tacrolimus is more potent than pimecrolimus. An in vitro study showed that T cell inhibition by tacrolimus was 8- fold higher than that with pimecrolimus.
They also point out that tacrolimus and pimecrolimus inhibit T lymphocyte activation.(They bind to an intracellular protein, FKBP12, which results in inhibition of calcineurin and prevents the activation of cytosolic nuclear factor of T lymphocytes. This effect reduces T lymphocyte proliferation and the release of proinflammatory cytokines. In addition, these drugs also prevent the release of mediators from mast cells.) These resulting immunosuppressive properties help in alleviating the skin inflammation found in atopic dermatitis but may also be related to cancer development.
With respect to children, the authors stated that: “younger subjects with a higher body surface area per weight and subjects with abnormal epidermis can have significant percutaneous absorption of topically applied calcineurin inhibitors, which may result in systemic serum concentrations known to cause immunosuppression.”
2. Background information about animal carcinogenicity of these drugs and earlier FDA concerns
FDA reviews of animal carcinogenicity and other studies on these drugs in 2000, before their approval stated the following :
The most striking and biologically relevant finding in the mouse dermal carcinogenicity study conducted with tacrolimus ointment was an increase in lymphoma. A statistically significant elevation in the incidence of pleomorphic lymphoma in high dose male (25/50) and female animals (27/50) and in the incidence of undifferentiated lymphoma in high dose female animals (13/50) was noted in the mouse dermal carcinogenicity study. This is the information that was recommended for inclusion in the Protopic lable in my original NDA review. It is still my opinion that this is the most significant tumor finding for tacrolimus ointment. It is not recommended that the incidence rates for liver carcinoma, uterine leiomyoma or cervix stromal sarcoma be included in the label because their biological relevance is questionable.
FHI has performed both single and repeated dose topical studies with14C-tacrolimus ointment in rats (Protopic NDA 50-777, Sections 126.96.36.199 and 188.8.131.52, Volume 9). These data demonstrate that after a single application of ointment to intact and abraded skin, the concentration of tacrolimus in lymph (mesenteric node) was greater than in blood. After repeated application for up to 14 days, there was no evidence of an increase in tacrolimus concentration in lymph nodes or in blood.
The following information is from a presentation by Dr. Susan Cummins of the FDA before this committee in 2005:
- Tacrolimus Ointment
- Lymphoma signal in dermal mouse carcinogenicity study
- Pimecrolimus Cream
- Lymphoma signal in oral mouse carcinogenicity study
- Lymphoma signal in 13 week dermal mouse study (pimecrolimus dissolved in ethanol)
- Pimecrolimus Cream-other tumor signals:
- Benign thymoma in oral rat carcinogenicity study
- Follicular cell adenoma of the thyroid in dermal rat carcinogenicity study
- Oral Primate Carcinogenicity Study
- Strongly positive for lymphoma
- Dose-response effect
- Additional reports of cancer and other serious adverse events in children & adults
- Increasing use of pimecrolimus cream & tacrolimus ointment, including in children younger than 2 years
Other concerns voiced by FDA at the 2005 meeting included the following by Dr. Dianne Murphy of the FDA:
- Many species, including non-human primates, developed cancers
- Biologic Plausibility very high
- Dose effect clearly seen in monkey study
- Documented high systemic levels in some cases after topical application
- Children have larger surface area and less evolved immune status
Dr. Nikhar of the Dermatologic Drugs Division of FDA showed a slide at the 2/05 meeting with the following information concerning systemic immunosuppression and lymphoma:
- Post transplant lymphoproliferative disorder (PTLD) in immunosuppressed patients related to Epstein-Barr virus infection is a well-recognized complication
- Risk of PTLD appears greatest in young children who are at risk for EBV infection while immunosuppressed*
- Risk appears to be related to the intensity and duration of immunosuppression
* Prograf label
3. New Safety Information presented by the FDA for this meeting
At the 2005 Pediatric Drugs Advisory Committee meeting, according to the FDA, there were 6 reports of malignancy in children through the age of 16. As of now, including the 10 additional cases of malignancy with these drugs in the FDA addendum, there are a total of 56 cases of malignancy in children up through the age of 16 known to the FDA. Of these, 16 were lymphomas. Since this meeting is focused on children, the FDA is not reporting the total malignancies with these drugs in people of all ages. (At the 2005 meeting, however, there were a total of 30 reports of malignancy for the two drugs, including the six in children mentioned above.)
It is of particular interest that four of the cases in children are T-cell lymphomas, in children ages 1.5, 6, 12 and 15. The 12 year old had a T-cell lymphoma at the application site of the drug. The FDA reports that NCI SEER data shows that 15.5% of cancers in children under the age of 20 are lymphomas in contrast to FDA reports that 36% (20) of the 56 malignancies reported to the agency were lymphomas, mainly non-Hodgkins (17 cases). Thus, lymphomas are more than twice over-represented in the FDA cases with these drugs compared to the data in the national SEER registry.
As will be presented by Dr. Spector shortly, 36% of the SEER lymphoma cases in children less than 20 are non-Hodgkins whereas 17/20 of the FDA lymphoma cases or 85% are non-Hodgkins lymphoma.
4. Actions to be Taken
When considering causality, five principles are mentioned as increasing the likelihood that a given cause (drug, in this case) is responsible for a specific effect:
a/ temporality: In the epidemiologic study cited above, the exposure to tacrolimus or pimecrolimus preceded the evolution of T-cell lymphoma
b/ biologic plausibility: these drugs target T-cells, are found in higher concentrations in lymph nodes than in blood and have clearly been shown to cause lymphomas in several species of animals
c/ specificity: the most statistically significant effect was seen in the increase in T-cell lymphomas
d/ strength of association: the hazard ratio for tacrolimus for T-cell lymphomas was 5.44 for tacrolimus and 2.33 for pimecrolimus
e/ dose effect: As mentioned by FDA’s Dr. Diane Murphy before this committee in 2005, there was a dose response relationship to lymphomas in the monkey study
It is thus highly likely, if not certain, that these drugs cause T-cell lymphomas in people just as they were found to increase the incidence of lymphomas in several species of animals.
In materials prepared for this meeting, the FDA stated:
“We also observed that, despite the addition of a Boxed Warning, both PIM and tTAC continue to be used outside the FDA approved labeling. Although the Boxed Warning states that CUI should not be used in children less than two years old, 50% of the malignancy cases and 32% of the infection cases reported starting the CUI in children less than two years old. The Boxed Warning for tTAC also states that only the 0.03% concentration is approved for children less than 16 years old; however, 26% of the malignancy and 22% of the infection cases reported the use of 0.1% tTAC. Finally, the Warnings section of the labels state that safety beyond one year of use has not been established; however, 41% of the malignancy cases and 24% of the infection cases that with information regarding the duration of therapy reported use for over one year.”
It is clear that although there has been a sharp decrease in the prescribing of these two topical drugs since the warnings in 2006, large numbers of children are still using the drugs. According to FDA data, there were approximately 30,000 children under the age of 2 were prescribed the two drugs in 2008-prescriptions for tacrolimus in children less than 2 actually increased from 2008 to 2009. In addition, according to FDA data, there were a total of 320, 000 children up to the age of 16 prescribed these drugs in 2008.
The box warnings and MedGuide have decreased prescribing for these drugs but almost 1/3 of a million children got prescriptions for them in 2008.
When used systemically, as in the oral or injected form, ProGraf (tacrolimus) appears to have benefits (suppressing organ transplant rejection) that outweigh the risks of cancer documented to occur with the drug. Thus, there is a favorable risk/benefit ratio.
The same cannot be said when the benefit is treatment of bothersome but not life-threatening eczema or atopic dermatitis. Giving children a drug known to increase the risk of T-cell lymphoma (and, possibly, other cancers) is a risk not worth taking for these clinical benefits. Rather than continuing on the spectrum of strengthening warnings for these topical drugs, the risk/benefit balance is such that their use should be prohibited in children. It is entirely possible that when the updated data on malignancies, especially lymphomas, in adults is obtained, our recommendation will also be to ban the use entirely for dermatologic purposes.
Thank you for your attention.