We describe the first documented case of an allergic reaction to the metacresol component of insulin in the pediatric T1DM literature. Allergy to the metacresol in insulin has been described previously in an adult with T1DM . Because metacresol is universally present in all current insulin preparations, we believe it has been overlooked as a possible cause of insulin allergy in some past case reports. The specific nature of the skin reaction, in our case erythema, urticaria, pain, and excoriation with abrasion and/or epidermal separation within 5 minutes of injection, is unusual and not previously described. It is probable that the skin reaction represents a mixed allergic response with Type I and IV allergy components. The wider literature on metacresol describes burn and skin breakdown reactions with cutaneous exposure , which may help explain this phenomenon. Metacresol is present in a variety of common products ranging from soaps to adhesives, agents to which our patient had experienced mild reactions in the past. A dose-response relationship was also observed: the lowest reaction was seen with Humulin NPH® (metacresol 1.6mg/mL) and the most severe reactions were seen with lispro and glulisine (metacresol 3.15mg/mL).
In the past, the diagnosis of insulin allergy was facilitated by commercially available insulin allergy test kits . These kits contained all the potential ingredients of commercial insulin preparations including several types of insulin and a range of excipients that included preservatives (e.g. metacresol), retardants (e.g. protamine sulfate), stabilizers (e.g. zinc), acid and base buffers, and isotonic agents (e.g. glycerol). Recently, manufacture of these kits has ceased with the current literature not responding to this change in circumstance.
We describe a novel approach to this dilemma. Common preparations are available that can be used via subcutaneous testing, in conjunction with blood tests, to isolate a specific cause of insulin allergy after excluding other common causes, for example injection technique, reaction to latex, skin preparations or needle types. The procedures to start a specific investigation are twofold: blood tests, including total IgE and specific IgE to human, porcine and bovine insulin as well as latex; and subcutaneous testing of all available insulin preparations. The taking of antihistamines during the testing and three days prior to testing is to be avoided. Because the active ingredient (insulin) and some excipients do differ between preparations, the pattern of reaction might suggest the cause. Next, Novo Nordisk™ diluting fluid can be helpful to distinguish an excipient allergy from true insulin and/or aspart allergy. Because the Novo Nordisk™ diluting fluid does not contain insulin but does have similar excipients to insulin aspart, a negative result would strongly suggest an allergic reaction to aspart and not to an excipient, whereas a positive result would indicate an excipient allergy. Investigating metacresol, the preservative universally present in available commercial insulin, is the next step. Lilly™ ‘saline’ for practice pen injection contains metacresol as the only ingredient other than sodium chloride, water for injection, and acid/base buffers. A positive reaction to subcutaneous testing with Lilly™ ‘saline’ strongly suggests metacresol allergy.
If diagnosed, options for treatment of non-spontaneously resolving metacresol allergy are few. Past insulin preparations did not contain metacresol, for example some porcine insulins, Monotard®, and Ultratard®. However, metacresol is present as a preservative in all currently available insulin, making a desensitization approach essential. Traditionally, successive subcutaneous injections of increasingly less dilute preparations of insulin are given . More recently, SCII has been used with success. In previous case reports, [6, 10] a low basal rate of between 0.1 units/hour and 0.3 units/hour has been successful at initiating desensitization; however, as we describe, substantially lower concentrations or even dilution with sterile saline may be required.
The time required for successful desensitization varies, and is patient and technique specific. If titration proves slow, then ketoacidosis from insulin deficiency quickly results. Intravenous insulin as previously described [10, 11] can provide an avenue for temporary insulin replacement during desensitization. Why intravenous therapy as opposed to subcutaneous therapy is generally well tolerated is not fully understood. Suggested mechanisms range from the simple mechanics of putting small volumes of insulin into a large central vein with subsequent rapid distribution (particularly relevant for localized reactions) through to differences in the immune system response depending on the route of insulin administration . Complications can still occur, with transient urticaria documented  along with the novel central chest pain seen in our patient.
Systemic treatment with oral antihistamine and/or steroid has been used . We found twice-daily oral antihistamine provided minimal improvement. Oral steroids possibly provided some benefit during our desensitization therapy, but are not ideal as a long-term option. Our patient continues on an antihistamine, but for subsequent local flare-ups we have had success using soluble hydrocortisone 0.1mL (50mg/mL) added to 1.9mL aspart insulin in the pump reservoir, to provide low-level local immune suppression.