Intended for healthcare professionals

Education And Debate

Lesson of the Week: Insulin dependent diabetes in nonagenarians

BMJ 1995; 310 doi: (Published 29 April 1995) Cite this as: BMJ 1995;310:1117
  1. Nigel D C Sturrock, registrara,
  2. Simon R Page, senior registrara,
  3. Pat Clarke, diabetes nurse specialista,
  4. Robert B Tattersall, professora
  1. a Diabetes Unit, University Hospital, Queen's Medical Centre, Nottingham NG7 2UH
  1. Correspondence to: Professor Tattersall.
  • Accepted 6 June 1994

The National Diabetes Data Group's definition of insulin dependent diabetes misleadingly talks of “predominant onset in youth.”1 Over 40 years ago Lawrence wrote that many diabetic patients aged over 70 were insulin deficient.2 It has recently been shown in Denmark that the annual incidence of insulin dependent diabetes after the age of 30 is 8.2 per 100 000 and is remarkably constant from the fourth to the ninth decade; thus at least 44% of all cases are diagnosed after the age of 30.3 We report on two patients who developed insulin dependent diabetes at the ages of 88 and 93.

Case reports


An 88 year old woman presented with a five month history of lethargy, weight loss, and nocturia. She was found to be hypothyroid, with a high titre of thyroid autoantibodies. One of her seven siblings had developed diabetes of uncertain type at the age of 75 and a granddaughter had Graves' disease. A month later she became thirsty, was found to have diabetes, and was treated with glibenclamide 5 mg twice daily by her general practitioner. She was referred to us three months later with a random blood glucose concentration of 15.1 mmol/l and a haemoglobin A1 concentration of 11.6% (normal range 5.0%-7.5%). Apart from bilateral cataracts there were no diabetic complications, but she had a body mass index of 23 kg/m2 and was positive for islet cell antibodies. A senior registrar diagnosed late onset insulin dependent diabetes but continued treatment with oral hypoglycaemic agents.

Over the next three years she had persistent hyperglycaemia but remained well and regularly walked about 3 km. Her weight gradually decreased, however, from 62 kg to 51 kg (body mass index 19.0) and, at the age of 92, she started taking Human Ultratard insulin 20 units once daily, the injections being given by her daughter. Within three months she had regained all her lost weight and was described as being like “a new person.” One year after taking insulin her treatment was changed to bovine lente insulin because of nocturnal hypoglycaemia. At the age of 97 she could still walk to the shops, and in five years of taking insulin she had not required admission to hospital or developed any diabetic complications. Paradoxically, diabetic control has been worse while taking insulin (haemoglobin A1 concentrations being in the range 12-14%), probably because the diabetes nurse allows her to have chocolate biscuits with her coffee and banana sandwiches for lunch, which according to the patient “make life worth living.”


A 93 year old woman with no family history of diabetes presented to a geriatric unit with a one week history of thirst and polyuria. In the 24 hours before admission she had become increasingly drowsy and confused. She had been wheelchair bound for four years because of osteoarthritis but had no other illnesses and was taking no drugs. She was dehydrated and cachectic with a Glasgow coma score of 3/14. She had moderate ketonuria, a blood glucose concentration of 44 mmol/l, and a serum osmolality of 396 mmol/kg. Blood gas pressures and bicarbonate concentration were not measured. Blood cultures, a chest radiograph, and an electrocardiogram showed nothing abnormal. Hyperosmolar non-ketotic coma was incorrectly diagnosed. After treatment with intravenous insulin and fluids she recovered full consciousness within 24 hours. Because she was thought to have non-insulin dependent diabetes she was discharged taking gliclazide 120 mg twice daily.

Over the next three months she did not see her general practitioner but had persistent symptoms, lost weight, and never felt well. She was then readmitted with heavy ketonuria, a blood glucose concentration of 31.3 mmol/l, and a blood bicarbonate concentration of 13 mmol/l. She was positive for islet cell antibodies at a titre of 1:64. Twice daily Human Mixtard insulin produced a rapid and sustained improvement in her symptoms and general wellbeing. The insulin was given by her daughter, with a district nurse providing holiday cover. Two years later she was well, had had no acute diabetic problems, and had no microvascular complications.


Insulin dependent diabetes in elderly people is commonly diagnosed insufficiently early to prevent prolonged ill health, futile treatment with oral hypoglycaemic agents, or a potentially fatal admission with ketoacidosis.4 Insulin dependent diabetes can be diagnosed by clinical criteria. Kilvert et al5 found that diabetes in patients aged 64 years or over who needed insulin within the first year was distinguished by ketonuria at diagnosis, a history of autoimmune endocrine disease, and low body weight. We previously found that low body weight or appreciable weight loss, combined with moderate or heavy ketonuria, was an accurate indication of insulin dependent diabetes at any age.1

The first patient had a normal body mass index, with subsequent appreciable weight loss despite maintaining reasonable glycaemic control. She also had autoimmune thyroid disease. In the second patient the onset was more dramatic: at presentation she was comatose with extreme hyperglycaemia and ketonuria. We believe that it is wrong to treat this type of patient without insulin irrespective of the islet cell antibody result since the mortality of recurrent hyperglycaemic coma in elderly people is high.6 Furthermore, elderly patients should not be allowed to languish in ill health when insulin could provide rapid relief of symptoms. Often the reluctance to start treatment with insulin in elderly patients arises from fears about possible hypoglycaemia and patients' ability to give injections. Our cases show that this should not be a problem: a relative or member of the nursing home staff can usually give it. If the patient can give the injection but not draw it up (a common problem in older people) a district nurse or relative can draw up a week's supply of syringes and leave them in the fridge.

If the aim of treatment is to control symptoms and restore vitality a once daily injection is usually adequate. It should nevertheless be noted that, as in case 1, human insulin is usually less satisfactory than bovine. Tindall et al found that, compared with bovine lente insulin, Humulin Zn (human insulin zinc suspension, Lilly) produced excessively frequent nocturnal hypoglycaemia.7

Our cases raise the interesting question of whether late onset insulin dependent diabetes has the same pathogenesis as that developing in childhood. If it has, why were our patients “protected” until their old age. Our patients were not tissue typed, but in Denmark 94% of elderly patients with presumed insulin dependent diabetes were positive for HLA DR3 or DR4, or both, a proportion identical with that in young patients.3 Furthermore, the prevalence of islet cell antibodies in these elderly Danish patients was only slightly lower than that in patients under 30. It therefore seems that the underlying process of ß cell destruction is the same irrespective of age at onset.

We thank Dr Jorn Nerup for letting us see a copy of his group's paper.3


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
View Abstract