Lack of Detectable Serum Levels Following Topical Fluoxetine Administration in a Child

To the Editor:

An 8-year-old boy with a history of attention-deficit/hyperactivity disorder (ADHD) combined subtype, complex partial epilepsy, pediatric migraine, nonverbal learning disorder, and anxiety disorder not otherwise specified was admitted to the inpatient child psychiatry unit of British Columbia's Children's Hospital in September 2012 for assessment and treatment. Admission medications included valproic acid (liquid) 250 mg twice daily, lisdexamfetamine 20 mg daily (dispersed in liquid), melatonin (sublingual) 5 mg at bedtime as needed, and fluoxetine 20 mg daily administered via topical application to alternating wrists every morning. The patient was extremely fearful of taking oral medications. This aversion developed after the family dog was euthanized. The patient feared that he would also die if he took oral medications. His parents had been restraining him and forcing him to take medically necessary oral medications, such as valproic acid for seizure prophylaxis.

Upon learning of the topical fluoxetine prescription, the clinical pharmacist contacted the dispensing community pharmacy for details of the product. Topical fluoxetine had been dispensed since February 2012 for ∼7 months prior to admission. The product was compounded to contain fluoxetine 100 mg/mL in Lipoderm^®, a proprietary permeation enhancer produced by the Pharmaceutical Compounding Centers of America.

A literature search for data regarding topical absorption of fluoxetine in humans did not reveal any relevant articles. Data obtained from veterinary literature indicated that transdermal fluoxetine absorption was likely to be low (Ciribassi et al. 2003; Trepanier 2011). The patient's mother described a pattern of response consistent with what may be typically expected following administration of oral fluoxetine for treatment of anxiety, including a reduction in symptoms that was delayed in onset by 3–4 weeks following the start of topical fluoxetine administration. No adverse effects were reported. The clinical pharmacist recommended checking a serum fluoxetine level to assess whether any of the fluoxetine was reaching the systemic circulation. The patient was also very fearful of needles, and the family initially refused to consent to having bloodwork drawn. The patient's family did not want to stop topical fluoxetine administration, even following a conversation with the treating psychiatrist and clinical pharmacist about the likely low rate of transdermal absorption. Continued use of this product was considered to likely be ineffective for treatment of the patient's anxiety disorder, but very unlikely to be directly harmful to the patient.

During the 1 month admission he was switched from liquid valproic acid to a divalproex sprinkle formulation (obtained via the Health Canada Special Access Program) which was consistently administered in hospital. With teaching and a rewards program implemented by unit staff, the patient became more amenable to taking medications orally. He no longer had to be restrained by his parents for medication administration. His parents consented to bloodwork to assess serum valproate and fluoxetine levels.

Consistent topical application of the fluoxetine product to alternate wrists during the month-long admission prior to the serum level draw was confirmed via the patient's medication administration records. Bloodwork was drawn on day 30 of admission at 0850 h, just prior to the scheduled morning topical fluoxetine administration time. The fluoxetine level and norfluoxetine (metabolite) level result were both reported as “not detected” (reference range: fluoxetine 300–1300 nmol/L, norfluoxetine 0–1200 nmol/L).

The fluoxetine/norfluoxetine level results were made available to the patient's community pediatrician following discharge with a request to re-evaluate the prescription in view of the undetectable serum levels.

Discussion

Transdermal medications absorb slowly from a depot formed in the skin following application (Trepanier 2011). Although administration via this route may provide a longer duration of action, it is possible that drugs may never reach therapeutic plasma concentrations (Trepanier 2011). No useful guideline is available to allow extrapolation of what the required transdermal dosage may be in comparison with the usual oral dosing (Trepanier 2011).

Although no research exists in humans, pharmacokinetics of topical fluoxetine have been assessed in the veterinary literature. The relative bioavailability of topically applied fluoxetine in permeation enhancing pluronic lecithin organogel (PLO) was determined to be ∼10% of orally administered fluoxetine in 12 healthy cats. Slower absorption and lower peak serum levels were noted with topical administration compared with oral administration (Ciribassi et al. 2003).

Our patient's topical fluoxetine product was compounded in Lipoderm. Lipoderm is a proprietary formula permeation enhancer, but is known to contain lecithin, and appears to have similar properties to those of PLO. Lipoderm has been shown to deliver gabapentin, clonidine, ketamine, and baclofen successfully through human cadaver skin in vitro (Pharmaceutical Compounding Centers of America 2013).

Fluoxetine is a small molecule (molecular weight 309.33 g/mol) and highly lipophilic (Hiemke and Härtter 2000), which are conditions thought to favor transdermal drug absorption (Trepanier 2011). In patients with normal hepatic function, fluoxetine half-life following chronic dosing is 4–6 days, whereas the norfluoxetine metabolite half-life is 4–16 days (Lexi-Comp 2013). Given this prolonged half-life, it would be expected that one would observe detectable fluoxetine levels following topical administration if an adequate amount of drug was reaching the systemic circulation. The lower detection limit of fluoxetine and norfluoxetine is reported as 40 nmol/L (Nichols et al. 1994). This is approximately one eighth of the lower limit of the fluoxetine reference range reported by our laboratory. Even if absorption of the topical dose was ∼10% of the expected absorption following an oral fluoxetine dose (as suggested by veterinary literature), serum fluoxetine levels should have been detectable if absorption was occurring.

In the cat study, topical fluoxetine was applied to the hairless portion of the pinnae (outer ear). In cats, this application site has relatively low skin thickness, and was selected “on the basis of being mostly hairless and easy to access” (Ciribassi et al. 2003). The stratum corneum (outermost layer of skin) is the principal barrier to transdermal drug absorption. Drug must diffuse through it and viable epidermis to reach the peripheral microcirculation and achieve systemic absorption and distribution (Sharma et al. 2011).

It is possible in humans that the wrist is not the optimal application site for transdermal drug absorption. In general, children's skin is more permeable than that of adults (Sharma et al. 2011), and this would be expected to promote drug absorption. In humans, the wrist (distal end of the volar forearm) has among the lowest stratum corneum thickness (16 μm), comparable to the abdomen (15 μm), forehead (13 μm) and back (10 μm).

Skin thickness and keratinization differs among species (Sharma et al. 2011) and may affect drug penetration. Results regarding transdermal drug absorption obtained in other species are not necessarily generalizable to humans, and human studies should be undertaken prior to clinical use of transdermal medications (especially outside of a palliative care setting).

Conclusion

Sustained daily topical application of fluoxetine 20 mg (compounded in Lipoderm) to the wrists of an 8-year-old boy failed to result in a detectable serum fluoxetine or norfluoxetine level.