Lipoatrophy as a cutaneous complication of insulin therapy has been extremely rare since the introduction of recombinant human insulin. Lispro insulin was not reported to be associated with this complication, but recently, Griffin et al. (1) published the first two cases of lipoatrophy associated with lispro insulin in two insulin pump-treated patients. We observed a singular case in which lipoatrophy occurred in two different locations with both buffered human regular insulin and lispro insulin in a patient treated by continuous subcutaneous insulin infusion (CSII).
We studied a 29-year-old Caucasian woman diagnosed with type 1 diabetes at 23 years of age. She began intensified insulin therapy with two premixed insulin injections before breakfast and dinner and regular insulin before lunch (22 units/day, 64% NPH insulin). She developed a primary hypothyroidism 1 year later and started treatment with sodium levothyroxine. She was transferred to our hospital in 1998. In the following 2 years, several combinations of NPH and regular insulin were used to reduce hypoglycemia, but optimized blood glucose was not achieved.
She began CSII (MiniMed 507 C; Soft-set Micro) with human regular insulin (Actrapid; NovoNordisk) in December 1999. Under CSII, HbA1c improved to 6.6–6.8% and the number of hypoglycemic episodes were reduced. Seven months after initiation of CSII, lipoatrophy was noted in the subcutaneous tissue of the anterior abdominal wall (Fig. 1A), although frequent change of catheter site every 3 days was performed. Initially, she was prompted to avoid the abdominal area for catheter insertion, but therapy was continued with buffered human regular insulin.
Treatment was also changed from buffered regular human insulin to lispro insulin 10 months later, due to deterioration of metabolic control. Eleven months after the initiation of lispro and using another 6-mm Teflon catheter (Quick-set), she noticed a new area of lipoatrophy on the right buttock, but it was of lesser extension (Fig. 1B). Both areas of lipoatrophy persisted until the date of submission for this article, and no further progression of skin lesions has been seen.
This case confirms a previous observation that lispro insulin might also induce lipoatrophy (1). Interestingly, lipoatrophy in our patient was associated with both lispro and recombinant human insulin. Although the lispro-induced lipoatrophic area was smaller, this case pointed out that local factors related to CSII use might play also a role. Nowadays, teflon catheters are the most common infusion sets used by pump-treated patients. They are placed by means of specially-designed devices. Although teflon cannulae are soft and comfortable, they could also induce local changes that lead to lipoatrophy in susceptible patients.
It has been previously suggested that lipoatrophy results from a local immune reaction to insulin crystals (2). The inflammatory response includes local hyperproduction of tumor necrosis factor-α from macrophages that led to differentiation of adipocites (lipoblastoma-like lipoatrophy) (2). Because lispro and recombinant human insulin had similar immunogenicity (3), the exact role of insulin autoantibodies is unclear. The reason that the lispro-induced lipoatrophic area in our patient was smaller may be related to the reduced ability of lispro insulin to aggregate.
Address correspondence to F. Javier Ampudia-Blasco, Diabetes Reference Unit, Endocrinology Department, Valencia Clinic University Hospital, Avda. Blasco Ibáñez, 17, E-46010 Valencia, Spain. E-mail: email@example.com.