Lipoatrophy Associated With Lispro Insulin in Insulin Pump Therapy
An old complication, a new cause?
- From the Department of Pediatrics (M.E.G., W.V.T.), School of Medicine, Yale University, New Haven, Connecticut; and the West Virginia School of Medicine (A.F.), Wheeling, West Virginia.
- Address correspondence to Margaret E. Griffin, MD, Pediatric Endocrinology, Yale University, 333 Cedar St., New Haven, CT 06520; e-mail: .
Lipoatrophy virtually disappeared as a cutaneous complication of insulin therapy after the introduction of recombinant human insulin. Consequently, we were surprised to observe two cases of marked lipoatrophy occurring in patients treated by lispro insulin administered by continuous subcutaneous insulin infusion (CSII).
The first case is a 10-year-old Caucasian girl diagnosed with type 1 diabetes at 4 years of age. HbA1c levels had ranged between 7.8 and 8.7% (normal ≤6.3%) on two daily injections of human NPH and regular insulin (Humulin-Regular; Eli Lilly, Indianapolis, IN) with ultralente (Eli Lilly) added subsequently. Apart from mild lipohypertrophy of the biceps area, there were no cutaneous complications. At 8 years of age, increased HbA1c levels and several severe hypoglycemic episodes prompted a switch to CSII (MiniMed model 507) using lispro insulin (Humulog; Eli Lilly), 35-40 U daily, 60% as basal replacement. HbA1c levels fell to 6.5-7.2%, and there were no further episodes of hypoglycemia. Twelve months after commencing pump therapy, lipoatrophy was noted in the subcutaneous tissues of the anterior abdominal wall and progressed over the next few months. Treatment was changed from lispro to buffered human regular insulin (Velosulin; Novo Nordisk, Princeton, NJ). There has been no further progression of skin lesions, although the lipoatrophic areas have persisted (Fig. 1).
The second case is a 51-year-old Caucasian woman who was diagnosed with type 1 diabetes at 12 years of age and began CSII with buffered human regular insulin (Velosulin; Novo Nordisk) in 1996 (HbA1c 8.1%). Previous treatment with prebreakfast injection of beef-pork NPH and regular insulin had resulted in no cutaneous complications. Ischemic heart disease, elevated LDL cholesterol, primary hypothyroidism, background of mild retinopathy, and distal sensory neuropathy were present. In 1998, treatment was changed to lispro insulin (Humulog; Eli Lilly) and HbA1c levels were between 5.5 and 6.4%. In the summer of 1999, she noticed lipoatrophy in the abdomen and thigh, and her bolus doses before meals were taking longer to peak, even when bolus was administered into nonaffected areas. Examination revealed lipoatrophy involving the abdominal wall, lateral thighs, and buttocks, all of which were sites of previous insulin infusions.
These cases highlight a potential for lispro insulin to induce lipoatrophy. While CSII may have contributed to the problem, use of lispro appears to be the most important factor. No further progression in lipoatrophy was noted in either patient after the switch to buffered human regular insulin using the same MiniMed pump systems. Moreover, one author (W.V.T.) met a teenage girl at a symposium who developed lipoatrophy with lispro delivered by a Disetronic pump.
Lipoatrophy in the era of less highly purified insulins was considered to have an immunologic basis. However, Fineberg et al. (1) have reported no differences in immunogenicity between lispro and recombinant human insulin. Nevertheless, lispro has proved to be an effective substitute for human regular insulin in several cases of presumed immunogenic insulin resistance (2,3,4). Thus, the mechanism causing lipoatrophy in our patients is unclear.
The purpose of this letter is to alert clinicians to a potential complication of lispro insulin not previously published or reported to Eli Lilly during premarketing or postmarketing studies (J. Holcombe, personal communication). Fortunately, this adverse effect appears to be uncommon and readily managed by switching to human regular insulin. It would be important to determine whether others have observed similar problems and whether these problems are limited to CSII.