We would like to add our experience of the clinical course and prognosis in patients with type 1 diabetes and eating disorders to that reported by Nielsen et al. (1) who reported a standardized mortality ratio of 14.5 in female patients with a diagnosis of type 1 diabetes and anorexia nervosa.
We performed a 12-year follow-up of a cohort of 14 women with type 1 diabetes and eating disorders (12 with anorexia nervosa) originally studied in 1987 (2). Of the 14 subjects, 5 had died (36%), 2 were blind, and 3 were receiving renal replacement therapy; most notably, 11 of the 14 subjects suffered from severe autonomic neuropathic symptoms. Data from 1999, or at the time of death, revealed that the median (range) age of the cohort was 37 years (25–46), with a duration of diabetes of 26 years (14–33). The age of death of the five patients was 30 years (25–42) with a duration of diabetes at death of 19 years (14–26).
Two patients were found dead at home and, as both had hypoglycemic unawareness, there was strong circumstantial evidence that the deaths were related to hypoglycemia. One patient died as a result of ketoacidosis after deliberate insulin omission. Another patient died following a sudden respiratory arrest 48 h after bone graft surgery for a nonuniting fracture. The final patient died from emaciation due to severe autonomic neuropathy affecting her bowel, resulting in intractable diarrhea and vomiting.
All 14 women had some degree of retinopathy and only 3 of the 14 women had normoalbuminuria. All but three women had complained of painful neuropathic symptoms. Interestingly, half of the cohort had at least one pregnancy, with one woman having three children.
Of the 14 women, 10 had recovered from their eating disorder in that they no longer fulfilled the criteria for diagnosis. Data from 1999, or at the time of death, revealed that the median (range) BMI of the cohort was 23.5 kg/m2 (16–30).
In summary, our results accord with those of Neilsen et al. We have found a 36% mortality over 12 years, and in keeping with other series, we report a high rate of microvascular morbidity (3,4).
Address correspondence to James D. Walker