We present the case of a 53 year old caucasian woman (R.P., born 1956), who was diagnosed with type-1 diabetes mellitus at the age of 13 years. Since 1984 she was treated by continuous subcutaneous insulin infusion with a portable insulin pump using human regular insulin, without any adverse effects. Since 2004, insulin lispro (Humalog®, Eli Lilly Deutschland, Bad Homburg, Germany) was used for insulin pump treatment (currently used device: Accu-Check Spirit®, Roche, Burgdorf, Switzerland). There was no co-morbidity except for Hashimoto's thyreoiditis, requiring 100 μg levothyroxine once daily for substitution. There was no evidence for diabetic retinopathy, nephropathy or neuropathy, respectively. HbA1c (normal range 4-6% of total haemoglobin) was around 8%, body mass index was 23.8 kg/m2. In November 2007 and February 2008, vaccinations against hepatitis A and B were carried out with a vaccine containing traces of thiomersal (Twinrix®, GlaxoSmithKline, Dresden, Germany). Commencing in May 2008, a lipoatrophic defect developed at the catheter insertion site on the right side of the abdominal wall, and after changing to the opposite side, a second lesion developed there. The size of both lesions increased steadily. In March 2009 hollows of 6 × 7 cm had developed in the subcutaneous fat tissue (Figure ). Routine blood chemistry was normal. Immunological serum markers (anti-insulin antibodies, as well as antibodies against mitochondria, c-ANCA, p-ANCA, and ANA IgG) were negative. Anti-TPO antibodies, however, were elevated (173 U/ml (normal < 60 U/ml)). In April 2009, a biopsy was obtained from the margin of the lipoatrophic area at the right lateral abdominal wall (Figure , lower panel). Routine histopathologic examination showed atrophic fat tissue (lipoatrophy); unfortunately, more detailed histochemical investigations were not performed. Assuming a potential adverse effect of Humalog®, the insulin preparation was changed-over to rapid-acting genetically engineered human insulin (Actrapid®, NovoNordisk, Mainz, Germany), without obvious benefit. In consideration of an undefined immunologic pathogenesis, low-dose immunosuppressive treatment with 10 mg prednisone once daily was commenced on June 23, 2009 (with adaptations of the insulin dosage, as appropriate, to maintain HbA1c between 7 and 8%). As an immediate effect, no new lipoatrophic hollows developed at the catheter insertion sites in use. In August 2009, the existing hollows started to re-fill, beginning at the margins (Figure , ). In September 2009, the daily prednisone dosage was reduced to 5 mg, and further to 2.5 mg in October 2009. In December 2009, the patient decided to switch back from human insulin (Actrapid®) to insulin lispro (Humalog®), while the lipoatrophic hollows continued to re-fill with subcutaneous fat tissue rather than fibrous scar tissue (confirmed by MRI, see Figure ). In March 2010, when the lipoatrophy was almost cured, prednisone was discontinued. The patient was advised to use the previously lipoatrophic areas again for catheter insertion. However, about 4 weeks later, prednisone had to be resumed because lipoatrophy had relapsed. By June 2010, after another 2 months of prednisone therapy, the lipoatrophic areas were cured again (Figure ) and prednisone was tapered off. During the following 12 months on unchanged insulin-pump therapy using insulin lispro, no relapse of the lipoatrophy was noted (Figure ).
Figure 1 Follow-up of lipoatrophic sites, before, during, and after prednisone therapy. Series of photographs showing lipoatrophic areas on both sides of the umbilicus (upper panel: left side, lower panel: right side). Photographs as of March 2009 and April 2009 (more ...)
Figure 2 Follow-up of lipoatrophic sites, before, during, and after prednisone therapy pt. 2. Continuation from Figure 1. Low-dose oral prednisone was administered from the last week of June 2009 until March 2010, and from May 2010 until June 2010. The photographs (more ...)
Figure 3 MRI demonstrating re-grown subcutaneous fat tissue at sites of previous lipoatrophy. Axial T1-weighted MRI sequence (1.5 Tesla Magnet, body surface coil, slice thickness 3 mm) acquired through the abdomen, March 2010. Markers are placed over the healed (more ...)
However, 3 months after cessation of prednisone treatment, in both hands cheiroarthropathy developed with thickened flexor digitorum tendons and edema of the tendon sheaths, as evidenced by MRI [15
]. Rheumatoid arthritis was ruled out on laboratory and imaging studies. The condition was markedly alleviated by physiotherapy.