1047
haplotype is different. In all, thirty families have now been reported from several centres and there has been no exception to this observation. Our findings also demonstrate a strong disequilibrium association with HLA antigen Bw47 (previously 407* CAS). The antigen Bw47 in this series was defined by our serum "CAS" (7W 387-workshop reference Bw47). Only in the group investigated by Zappacosta et al. has there been any association with a particular HLA antigen (Aw29, B35 in their series). This strong association between Bw47 and the abnormal gene has not been commented on by other investigators. Bw47 is considered to be uncommon in most White populations, with a frequency of 0.005 in our control population of 1000 normal blood-donors, compared with a frequency of 0.54 among the affected families (seven out of our series of thirteen families). In one family the patient is homozygous for Bw47. Two affected sibs homozygous for Bw47 were reported by the
Irvine’s is not clear. It is possible that the duration of diabetes is longer in our patients and that immune-complex concentrations decline with duration of disease and are related to islet-cell antibodies rather than insulin therapy. Alternatively, if complexes do consist of insulin and insulin antibody their concentration may fluctuate with plasma-insulin concentrations and be related to diabetic control. Department of Diabetes and Endocrinology, Dudley Road Hospital, Birmingham B18 7QH Department of Clinical
Pharmacology, University of Birmingham Department of Immunology, University of Birmingham
D. H. COVE K. L. WOODS H. M. CHAPEL C. L. HALL
Price et al .7
Although this association with Bw47 is interesting, in most families studied the combinations between HLA-B alleles and the abnormal gene are so variable that it is not possible to speculate about the association. A table V.P.
showing haplotypes in
all thirteen families may be had from
Genetic
Counselling Service, Department of Pædiatrics, City Hospital, Nottingham NG5
1PB
Tissue Typing Laboratory,
Regional Transfusion Centre, Sheffield
V. PUCHOLT
J. S. FITZSIMMONS K. GELSTHORPE R. F. PRATT R. W. DOUGHTY
IMMUNE COMPLEXES IN DIABETIC RETINOPATHY
SIR,-Vascular damage in a number of diseases is associated with the presence of circulating immune complexes.Dr Irvine (Sept. 2, p. 528) refers to an increased prevalence of immune complexes in diabetics with proliferative retinopathy compared with diabetics of a similar duration without retinopathy, and suggests that immune complexes may be important in the pathogenesis of diabetic microangiopathy. Using the Raji cell method2 we have measured immune complexes in the sera of 12 patients with proliferative retinopathy (group A) and 23 patients with minimal background retinopathy or no retinopathy (group B). All patients were insulin dependent and had received insulin for at least 10 years (mean duration 21.4 years in group A and 21.6 years in group B). The mean ages of the two groups were similar (group A 50 years, group B 48 years). Patients were receiving a variety of insulin regimens. Blood-samples were taken 5-8 h after their normal morning injection. One patient with proliferative retinopathy had a slightly raised serum concentration of immune complexes at 20 fLg/ml (normal < 16 .g/ml). All other concentrations were normal, and the mean concentrations were the same in group A (7-99 g/ml) and group B (7-8fLg/ml). (Results are expressed as equivalents of g of aggregated human gamma-globulin per
ml.) Thus serum immune complex concentrations are rarely increased in long-standing insulin-dependent diabetics, yet it is these patients who are at greatest risk of proliferative retinopathy and other features of diabetic microvascular disease. Our results suggest that immune complexes are unlikely to be important in the pathogenesis of proliferative retinopathy. The cause of the discrepancy between our results and 6. 7.
Zappacosta, S. and others ibid. 1978, ii, 524. Price, D. A., Klouda, P. T., Harris, R. ibid. 1978, i, 930. 1. Pussell, B. A., Lockwood, C. M., Scott, D. M., Pinching, A. J., Peters, D. K. Lancet, 1978, ii, 359. 2. Theofilopoulos, A. N., Wilson, C. B., Dixon, F. J. J. clin. Invest. 1976, 57, 169.
WHEAT-SENSITIVE—BUT NOT CŒLIAC
SIR,-A 16-month-old girl was referred for investigation because of diarrhoea, irritability, and loss of appetite. She was breast-fed for 9 months and then gradually advanced to a full diet. At 12 months she became irritable and had three or four loose, bulky stools each day. Recurrent upper respiratory infections and anorexia followed and she stopped gaining weight. Physical examination was unremarkable, apart from a slightly protruding abdomen. Stools were free from pathogenic bacteria or parasites. Stool pH was between 6-$—7-0. The coefficient of fat absorption was 92%. Total eosinophil count 900/1. Small intestinal biopsy, on a full diet, showed normal histology, including intraepithelial lymphocyte count and mucosal disaccharidase levels (maltase 182 u/g protein, sucrase 80 u/g, lactase 21 u/g). The mother had, in infancy, recurrent vomiting and diarrhoea related to gluten-containing food. Her father, a poediatrician, diagnosed coeliac disease; she had a strict glutenfree diet until the age of 4 years. She still avoids bread and pastry which cause abdominal pain and loose stools. Intestinal biopsy, after her daughter’s investigation, showed normal his-
tology. Because of this family history, and on the mother’s insisthat the child’s symptoms were related to wheat, we started the child on a gluten-free diet. Within days her symptoms subsided. Rechallenge with wheat induced diarrhoea within 24 h; irritability and anorexia followed within days. At present, at the age of 3 years, the child avoids wheat-containing foods by choice. We consider this patient’s symptoms to be due to wheat sensitivity, probably of an allergic nature. The normal jejunal histology and absorptive function exclude classical glutensensitive enteropathyl or transient gluten intolerance.2,3The clinical response to wheat withdrawal and rechallenge are, like milk allergy, frequently encountered in infancy.4 Children of atopic mothers have been considered predisposed to atopic diseases of the skin and respiratory and gastrointestinal tracts.5 Our patient’s family history suggests a strong hereditary tence
predisposition. Gluten-free diets for children with a variety of chronic gastrointestinal symptoms are, unfortunately, still prescribed without full investigation. Many diagnostic misinterpretations have followed this practice.6 A favourable response to such a diet does not necessarily imply cceliac disease, as this case
show. Pediatric Gastroenterology, Chaim Sheba Medical Centre, Tel-Hashomer, Israel
ANITA JONAS
1. Strober, W. Ann. intern. Med. 1975, 83, 242. 2. Walker-Smith, J. Archs Dis. Childh. 1970, 43, 523. 3. Visakorpi, J. K., Immonen, P. Acta pæd. scand. 1967, 56, 49. 4. Eastman, E. J. Pediatrics, 1977, 60, 477. 5. Dannaeus, A., Johannsson, S., Fouchard, T. Acta pœd. scand. 1978, 67, 497. 6. Lloyd-Still, J. D. Pediatrics, 1978, 61, 929.