Novel HLA-DPB1 alleles detected in the Ethiopian population.

The number of identified HLA-DPB1 alleles increased rapidly by application of DNA-based typing techniques. PCR-SSO typing indicated the presence of possible new HLA-DPB1 variants in the Ethiopian population. The use of the SBT technique, which considers polymorphic as well as constant regions in the second exon of HLA genes, allowed direct identification of two new allelic variants. Moreover, a recently identified HLA-DPA1 variant was also present in this population. The newly defined allelic HLA-DPB1 sequences found in five individuals of the Ethiopian population were confirmed by cloning and subsequent sequencing of the cloned DNA. One of the new alleles was shown to segregate in a family and was also present in unrelated individuals. Both new DPB1 alleles represent new combinations of existing polymorphism in the hypervariable regions. In different populations the frequency of these new HLA-DP variants remains to be determined.

Relative contribution of HLA-DQA and -DQB alleles to predisposition to insulin-dependent diabetes mellitus.

The presence of DQA and DQB alleles conferring protection or susceptibility was assessed in a panel of 39 insulin-dependent diabetes mellitus patients and 39 healthy control subjects from the central highland of Ethiopia. The results were grouped into three entities: a combination of alleles conferring susceptibility, a group conferring protection, and a group without any apparent HLA-DQ or -DR predisposition to insulin-dependent (type 1) diabetes mellitus (IDDM). Statistical analysis revealed that the relative risk of the first group is 64.1. If a similar approach is applied to the data on a study in caucasoid IDDM patients and controls of Kahlil and colleagues, the pattern is fully consistent with the data presented here, with an extraordinarily high relative risk (RR 258.2). It will be of interest to study whether this subdivision is reflected or supported by clinical or etiologic differences of the disease. The predictive value of susceptibility phenotypes appears to be more accurate by the proposed subdivision. Furthermore, in combination with islet-cell antibody analysis, assessment of genotype will permit more accurate identification of prediabetic individuals to be entered in clinical trials.

Low levels of serum calcidiol in an African population compared to a North European population.

OBJECTIVE: To compare vitamin D status in an African population living at 10 degrees N with a Norwegian population living at 60 degrees N. DESIGN: Serum samples from 30 healthy young Ethiopians and 31 full term pregnant women from Addis Ababa were collected in September, and from 24 healthy Norwegians in March and 23 pregnant women from Oslo in February to June. METHODS: Serum (s) levels of calcidiol and intact parathyroid hormone (iPTH) were measured. RESULTS: The median values for s-calcidiol were significantly lower in Ethiopians compared with Norwegians (young Ethiopians 23.5nmol/l vs young Norwegians 81nmol/l, P<0.001; pregnant Ethiopians 25nmol/l vs pregnant Norwegians 36nmol/l, P<0.05) while those for s-iPTH were significantly higher (young Ethiopians 5.7pmol/l vs young Norwegians 2.4pmol/l, P<0.001; pregnant Ethiopians 4.8pmol/l vs pregnant Norwegians 2.8pmol/l, P<0.02). CONCLUSION: In spite of abundant availability of ultraviolet radiation, the population from Addis Ababa had a high rate of biochemical vitamin D deficiency compared with the Norwegian group.

Thyrotoxicosis in Ethiopian patients--a prospective study.

Forty-six cases of thyrotoxicosis, 8 males and 38 females, were seen over a period of two years from March 1979. There were 19 cases of toxic multinodular goitre, 22 of Graves' disease, four of toxic adenoma and one of subacute thyroiditis. The median age of those with toxic multinodular goitre was 49 and of those with Graves' disease was 30 years. The five most common symptoms were sweating/heat intolerance (82.6%), weight loss (82.6%), palpitations (82.6%), nervousness and irritability (73.9%) and insomnia (69.6%). Six of 24 non-pregnant females of child-bearing age had menstrual disturbances. The five most common signs were goitre (97.8%), tachycardia (100.0%), sweating (84.7%), and fine finger and tongue tremors (89.1%). Among the 22 cases of Graves' disease, 13 (59.1%) had exophthalmos and 4 (18.2%) had pretibial myxoedema. Treatment with propylthiouracil or carbimazole was effective in controlling the thyrotoxicosis. Granulocytopenia developed in two cases and skin allergy in one other, necessitating use of alternative treatment. Seven cases had subtotal thyroidectomy after preparation with antithyroid drugs or propranolol and Lugol's iodine.

Insulin-dependent ketosis-resistant diabetes in Ethiopia.

Anthropometric, clinical and biochemical findings were compared in 30 rural (group A), 18 urban insulin-requiring (group B) and 45 urban oral-agent-responsive (group C) newly diagnosed diabetics. Mean ages at onset were 28.3 +/- 12.0, 25.6 +/- 14.5 and 42.1 +/- 10.5 years respectively. The differences between A and C and between B and C were significant. Group A were poor and malnourished, with body mass index (BMI) 15.9 +/- 1.9 and 17.2 +/- 3.7 kg/m2 for males and females respectively, presented with a long history of classical diabetes without ketoacidosis and required insulin in modest doses. 3 of 10 cases had excess stool fat but none of 13 unselected cases had pancreatic calcification. Group C were better nourished, with BMI 22.6 +/- 2.8 and 22.4 +/- 4.5 kg/m2, and responded to oral agents. Group B, with BMI 17.2 +/- 2.6 and 18.6 +/- 3.1 kg/m2, required insulin for control but had C-peptide levels above 0.02 nmol/1 in 10 of 15 cases. Anthropometric indices for males, but not for females, were significantly lower in group A than in group B or C. There were significant differences in levels of glucose between A and B and A and C, free fatty acids between A and C and B and C, insulin between A and B and A and C and C-peptide between A and C and B and C. Of the 3 groups the rural type most closely resembled the tropical variants.

Heritability of diabetes mellitus in Ethiopian diabetics.

A prospective case control study was conducted in 859 diabetic probands and 1059 nondiabetic controls. These were interviewed for history of diabetes mellitus among their first degree relatives. The protocol also included second and third degree relatives. The prevalence of diabetes mellitus in the first degree relatives of cases and controls showed a statistically significant difference (Z = 6.7564 and P = 0.00000). Similarly, the prevalence of diabetes among the second and third degree relatives of cases and controls differed significantly (Z = 2.74, P = 0.006). There were 445 non-insulin dependent diabetes mellitus (NIDDM) and 414 insulin dependent diabetes mellitus (IDDM) in the diabetic probands. There was more family history of diabetes mellitus among the first degree relatives of NIDDM than IDDM and the difference is statistically significant (Z = 4.076, P = 0000). The second and third degree relatives of NIDDM and IDDM also differed significantly (Z = 4.05, P = 00005). This study indicates that heredity plays an important role in the genesis of diabetes mellitus in Ethiopian diabetics.

Prevalence and clinical features of tuberculosis in Ethiopian diabetic patients.

OBJECTIVE: To determine the prevalence and clinical characteristics of tuberculosis (TB) in diabetic patients. DESIGN: This study was a cross-sectional survey based on the retrospective analysis of data on tuberculosis in diabetic patients. SETTINGS: The study was conducted at Endocrinology and Metabolism Unit of the Department of Internal Medicine, Faculty of Medicine, Addis Ababa, Ethiopia. SUBJECTS: Seventy-eight cases of tuberculosis among 1352 diabetic patients were included from September 1989 to 1996. MAIN OUTCOME MEASURES: Clinical evaluation, chest x-ray, acid fast bacilli (AFB) in sputum or measures tissue and histopathologic characteristic of biopsy specimens. RESULTS: Seventy-eight cases of TB were identified among 1352 diabetic patients giving a prevalence of 5.8%. Among 1352 diabetic patients, 713 (52.7%) were males, 639 (43.3%) were females, 619 (45.8%) were IDDM and 733 (54.2%) were NIDDM. The mean age of the 71 TB patients whose records could be retrieved was 34.0 +/- 11.9 years, 42(59%) were males, 29 (41%) were females. Fifty-four (76.1%) were IDDM and 17(23.9%) were NIDDM, 17 of the IDDM had clinical characteristics of malnutrition-related diabetes mellitus (MRDM), 36 (56%) of 71 patients were admitted for management. The three most common symptoms of tuberculosis were fever (80.5%), sweating (80.4%) and cough (70.5%). Twenty six (36.6%) of 71 cases were positive for AFB and six (8.5%) were seropositive for HIV. Fifty-six (78.9%) had pulmonary,eight (11.2%) had extrapulmonary and seven(9.8%) had disseminated TB. Forty-eight of 53 abnormal chest x-ray showed unilateral involvement. Thirty-eight of 41 (53.6%) had completed TB treatment, relapse occurred in seven (9.8%), eight (11.2%) are currently on treatment, 13 (18.3%) were lost to follow up, four (5.6%) defaulted and three (4.2%) died. The relative risk of developing TB in IDDM and NIDDM patients was being 26 times and seven times than the general population respectively. CONCLUSION: The prevalence of tuberculosis in the diabetic population is high and this warrants a prospective study to determine association between pulmonary tuberculosis and diabetes.

PIP: This retrospective, cross-sectional analysis examines the prevalence and clinical characteristics of tuberculosis (TB) in Ethiopian diabetic patients. Using the records of 1352 patients seen at the Endocrinology and Metabolism Unit of the Department of Internal Medicine, Faculty of Medicine, Addis Ababa, Ethiopia, 78 cases (5.8%) of TB were identified. Diagnosis of the disease includes clinical evaluations, chest x-ray, and detection of acid-fast bacilli (AFB) in the sputum or measures tissue and histopathologic characteristics of biopsy specimens. Prevalence was 52.7% among males, 54.25% among those with non-insulin-dependent diabetes mellitus (NIDDM), 47.3% among females, and 45.8% among those with insulin-dependent diabetes mellitus (IDDM). The mean age of 71 patients with TB was 34 +or- 19 years (14-85 years); 76.0% were IDDM and 23.9% were NIDDM patients. 17 of the IDDM patients also had clinical characteristics similar to those of malnutrition-related diabetes mellitus. The most frequent symptoms of TB included fever (80.5%), sweating (80.5%), and cough (76%); 78.9% of patients suffered from pulmonary, 11.2% from extrapulmonary, and 9.9% from disseminated TB. On the other hand, chest X-ray reports revealed that 48 of the patients had unilateral or multiple lobe involvement and 5 had bilateral involvement. Of the 71 patients, 38 were able to complete antituberculosis treatment, 4 patients defaulted, 7 had relapse, 6 were still in treatment, 13 were lost to follow-up, and 3 died. In conclusion, this study confirmed the 26- and 7-fold risk of developing TB among IDDM and NIDDM patients, respectively.

Neonatal congenital hypothyroidism screening in Addis Ababa, Ethiopia.

OBJECTIVES: To determine the prevalence of congenital hypothyroidism, establish reference levels for thyroid stimulating hormone (TSH), and determine the cut-off points of TSH level for neonatal screening in congenital hypothyroidism. DESIGN: A cross-sectional study. SETTING: Ghandhi Memorial, Zewditu, Tikur Anbessa and St Paulo's referral hospitals in Addis Ababa, Ethiopia, from July 1996 to January 1997. SUBJECTS: Four thousand two hundred and six consecutive newborns. MAIN OUTCOME MEASURES: TSH and T4 values using standard TSH level estimation approach. RESULTS: The overall mean TSH value was 9.6 mIU/l with standard deviation of 7.8 mIU/l. Transient hypothyroidism occurred in 3.6% of the neonates. No true case of congenital hypothyroidism was identified. The mean(SD) sampling time was 12.8(6.7) hours and the recall rate 4.8%. Most neonates (98.6%) were screened at or less than 24 hours of age. Twins had lower mean TSH value of 7.9(SD=6.4) mIU/l than single neonates with 9.6(SD=7.8) mIU/l, while females had significantly lower mean TSH value of 9.4(SD=7.9) mIU/l than males of 9.8(SD=7.6) mIU/l. The mean TSH level decreased significantly as sampling time increases, the mean TSH level for neonates older than 24 hours is significantly lower than the mean TSH level for those with 24 hours old or less. No significant difference was found in the mean TSH levels in relation to maternal age, parity and gestational age. No significant correlation was found between TSH value and birth weight (r=0.02), height (r=0.03), and head circumference (r=0.02) of the neonates. The most appropriate TSH level cut-off point was found to be at 29.4 mIU/l. The reference range (the range between 2.5th and 97.5th percentiles) for serum thyrotrophin value was higher in the age range of 6 to 12 hours and then declined. CONCLUSION: This pilot study has highlighted important findings, however larger studies should be conducted to assess the magnitude and associated factors of congenital hypothyroidism because congenital hypothyroidism represents one of the most common preventable causes of mental retardation.

Profile of coronary artery risk factors in Ethiopian diabetic patients.

OBJECTIVES: To assess the coronary artery risk factors in Ethiopian diabetic patients. DESIGN: A cross sectional study done on a representative sample of diabetic patients to evaluate the coronary risk factors. SETTING: Hospital-based study. SUBJECTS: A total of 302 diabetic patients randomly selected from the 3000 regularly attending diabetic patients in the Tikur Anbessa Hospital Diabetes Centre of whom 161(53.3%) were males, 141(46.7%) were females, 140(46.4%) were Type 1 and 162(53.6%) were Type 2. OUTCOME MEASURES: Assessment of the coronary risk factors in diabetic patients. RESULTS: There were 20(6.6%) smokers and no ex-smokers. All smokers were males of whom five (25%) were Type 1 and 15(75%) were Type 2. Hypertension and obesity were found in 64(21.2%) and 69(22.8%) patients respectively. Sixty one (20.2%) patients and 43(14.2%) patients had hypercholesterolaemia and hypertriglyceridaemia respectively. CONCLUSION: Our diabetic patients share the risk factors for developing coronary artery disease like any other diabetic patient in the developed countries but at a lower level.

Normal thyroid function values in Ethiopians.

Thyroid function values were determined in 56 healthy nongoitrous adult Ethiopians. The mean triiodothyronine (T3) values for 20 males and 36 females were 1.42 +/- 0.32 nmol/L and 1.51 +/- 0.25 nmol/L, and thyroxine (T4) values were 119 22 nmol/L and 116 +/- 21 nmol/L respectively. The mean thyrotropin (TSH) values for males and females were identical at 1.85 +/- 0.94 mu/L. Radioactive iodine uptake (RAIU) at 2 hours was 8.6 +/- 4.4% in males and 11.3 +/- 4.3% in females, and at 24 hours 31.7 +/- 11.7% and 38.9 +/- 11.1% respectively. The differences between males and females were significant at both 2 and 24 hours (P less than 0.05). The ranges for the 3 hormones derived from the mean 2SD values are close to those supplied with the kits but the Ethiopian RAIU values are higher than the values currently applicable in developed countries, probably indicative of the lower level of dietary iodine available to the population here. The values obtained in this study are offered to serve as normal reference for the interpretation of thyroid function results in Ethiopian patients. A strategy for the rational utilization of the available in vivo [corrected] tests is suggested.

Urine glucose testing: another look at its relevance when blood glucose monitoring is unaffordable.

The reliability of urine glucose testing to monitor diabetic control was investigated in patients attending the Diabetic Clinic of the Tikur Anbassa Hospital in Addis Ababa between October 1994 and January 1995 with the aim of utilising it for those with a normal renal threshold who cannot afford the cost of home blood glucose monitoring. Clinically important fasting blood glucose values were taken as those > 180 mg/dl and important urine glucose values as those > or = 0.25% by Clinitest. Capillary blood glucose was determined by visual and metre readings. Urine was tested for glucose by the standard Clinitest method. There were 265 patients, 126 IDDM and 139 NIDDM. Urine glucose corresponded satisfactorily with FBG in 80% of the patients. The sensitivity, specificity, positive and negative predictive values of urine glucose results by Clinitest in comparison with FBG by metre determination were 71%, 90%, 90% and 70% and by visual determination 71%, 86%, 80%, 79% respectively. On the basis of these results we conclude that urine glucose testing by Clinitest provides reliable information in more than 70% of our diabetic patients the majority of whom cannot afford the cost of home blood glucose monitoring.

A study of the transfer of insulin-requiring diabetic patients from U-40 to U-100 Lente insulin.

A pilot study of the transfer of insulin-requiring diabetic patients from U-40 to U-100 Lente insulin was conducted in 16 patients. The patients were returned to U-40 after 12 weeks on U-100. There were no adverse events specifically attributable to the process of transfer. The mean blood glucose values for the last day and the last 2 days on U-100 insulin (154.58 +/- 32.09 mg% and 152.99 +/- 29.23 mg% respectively) were significantly lower than those for the corresponding periods on U-40 insulin (178.06 +/- 38.48 mg% and 169.83 +/- 32.34 mg% respectively), (t = 2.79, p = 0.014 and t = 2.30, p = 0.036 respectively). But there was no significant difference between the mean values for the last but one day on U-40 and U-100 (152.36 +/- 27.81 vs 161.59 +/- 33.82; t = 1.16, p = 0.264). There was no significant increase in frequency (2.3 episodes per week on U40 versus 2.8 episodes per week on U-100) or severity of hypoglycaemic attacks with U-100 insulin. On the basis of this experience we feel that transfer from U-40 to U-100 insulin can be achieved without risk of serious adverse consequences provided the transfer is carried out according to a well-designed plan of action.

Radioiodine treatment of hyperthyroidism at the Tikur Anbessa Hospital.

Thirty-four hyperthyroid patients selected on preset criteria for radioiodine therapy were given low dose radioiodine-131 (I-131) in the Nuclear Medicine Unit of the Tikur Anbessa Hospital as a preferential alternative to repeat or protracted administration of antithyroid drugs, or surgery. The dose chosen was 85 uCi per gram of thyroid tissue based on the 24 hour uptake of a test dose of I-131. The age range of the patients was 18 to 68 years, female to male ratio was 7.5:1 and 80% of the females were in the reproductive age group. Twenty-five patients had multinodular goiter, four had persistent hyperthyroidism after surgery, and five had recurrent Graves' disease. Estimated thyroid gland weight was from 3.5 to 51.4 grams and the amount of I-131 administered was from 0.2 to 4.9 mCi. Only 11 of the 34 patents became euthyroid after a single dose of I-131, 14 required one or more repeat doses and only six of those became euthyroid making a total of 17 (50%) responders. The low dose regimen we used in this study, though economical in a situation of limited I-131 supply, resulted in a high failure rate of 50%. This shows the need to formulate a better dosage scheme that can benefit the majority of patients by optimizing the prospects of the resolution of hyperthyroidism with a single dose of radioiodine.

The relation of early nutrition, infections and socio-economic factors to the development of childhood diabetes.

The relationship of development of diabetes from birth up to 15 years of age to the type of feeding in infancy, childhood infections and vaccination was studied in 55 patients attending Endocrinology Clinics of the Ethio-Swedish Children's and Tikur Anbessa Hospitals over a period of two years (January 1990 to December 1991). Seventy-four unaffected siblings and 107 unrelated controls were interviewed for comparison. No significant difference was found in relation to type of feeding up to the ages of three, six and 12 months or older between patients and unaffected siblings. Histories of measles, chicken pox and whooping cough were equally distributed between the two groups. However, introduction of bottle-feeding was significantly more frequent among unrelated controls at three months of age (9/39 diabetics versus 41/83 controls) and six months of age (26/39 diabetics vs 72/83 controls) chi 2 = 6.6, (p < 0.01) and chi 2 = 5.6 (p < 0.05) respectively. The odds ratios between diabetics and unrelated controls for introduction of bottle-feeding at three months and six months of age were 0.32 (confidence intervals 0.14-0.74) and 0.31 (confidence intervals 0.13-0.77) respectively. The use of cow's milk and other formulas in bottle-feeding showed a significant negative association with the development of diabetes chi 2 = 5.8 (p < 0.025), chi 2 = 3.8 (p < 0.05) respectively. A history of vaccination against tuberculosis, measles, diphtheria, pertussis, tetanus (DPT) and polio was significantly more common among unrelated controls than diabetics, chi 2 from 4.6 to 11.4 (p < 0.05 to p < 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)

PIP: The development of insulin-dependent diabetes mellitus is thought to be related to environmental trigger factors acting upon a background of genetic predisposition to the disease. Viral infections, toxins, and dietary factors are being considered as trigger factors. The authors studied the relationship of the development of diabetes from birth up to 15 years of age to the type of feeding in infancy, childhood infections, and vaccination among 55 patients attending Endocrinology Clinics of the Ethio-Swedish Children's and Tikur Anbessa Hospitals over the period January 1990 to December 1991. 74 unaffected siblings and 107 unrelated individuals were included as controls. No significant difference was found in relation to type of feeding up to the ages of three, six, and twelve months or older between patients and unaffected siblings. Histories of measles, chicken pox, and whooping cough were equally distributed between the two groups. The introduction of bottle-feeding, however, was significantly more frequent among unrelated controls at three months and six months of age. The use of cow's milk and other formulas in bottle-feeding showed a significant negative association with the development of diabetes. An history of vaccination against tuberculosis, measles, diphtheria, pertussis, tetanus, and polio was significantly more common among unrelated controls than diabetics. There was no significant difference in family history of diabetes in first degree relatives, parental education, and level of income between diabetics and unrelated controls. The authors suggest that more extensive studies are warranted.

Light and electron microscopic changes in sural nerves in Ethiopian diabetics.

The light and electron microscopic morphological changes in sural nerve biopsies from 32 Ethiopian diabetics and 4 malnourished non-diabetic subjects were studied in order to determine the role of malnutrition in the development of diabetic neuropathy. The most prominent finding in diabetics with clinical neuropathy was reduction in the density of myelinated axons of all diameters especially the large ones. Affected nerves showed abundant intraneural collagen, and evidence of nerve regeneration in some cases. Demyelination, though present, was less marked. Changes in the vasa nervorum were seen in only one case. The morphological changes correlated with the duration and severity of the diabetes and with nerve conduction velocity. No changes were seen in malnourished non-diabetics and diabetics without clinical neuropathy, thus suggesting that malnutrition by itself does not contribute significantly to the development of diabetic neuropathy. Axonal degeneration appears to be the predominant feature in diabetic neuropathy.

Cushing's syndrome: a ten year experience at Tikur Anbassa Hospital.

A retrospective analysis of Cushing's syndrome from the retrieved records of 24 out of a total of 27 referred cases over a period of 10 years (January 1985-January 1995) showed confirmation in 16 cases. Fourteen (88%) were females, 9 (56%) were in their third decade. Duration of symptoms before admission was one year or less in eleven (69%) patients. Among symptoms weakness, fatigue and weight gain and among signs moon face, truncal obesity and buffalo hump were found in all patients. Amenorrhea occurred in all the women of childbearing age. Basal urinary 17-hydroxycorticoid (17-OHCS) values were elevated in 9/12 patients and low dose dexamethasone suppression test favoured Cushing's syndrome in 8/9 patients. Definitive treatment consisted of bilateral adrenalectomy in 7 and transfrontal pituitary adenomectomy in 3. Two patients declined surgery; 4 patients were lost to follow up before definitive treatment. Two patients who had bilateral adrenalectomy and two who had transfrontal adenomectomy died subsequently while three remained in good health on corticosteroid replacement therapy. Three of the 7 patients who had bilateral adrenalectomy developed Nelson's syndrome. It is concluded that Cushing's disease is the most frequent cause of Cushing's syndrome in our series. The patients presented with the well-known clinical manifestations of hypercortisolism. However, our laboratory facilities require considerable improvement to serve as a reliable adjunct to clinical evaluation. Since Cushing's syndrome, whatever the cause, is eventually fatal if left untreated, surgical intervention is mandatory after confirmation of the cause with appropriate laboratory tests.

Retinopathy in patients of Tikur Anbessa Hospital diabetic clinic.

A total of 302 diabetic patients were selected from regular attendants of the Tikur Anbessa Hospital (TAH) diabetic clinic to determine the prevalence of retinopathy from December 1994 to March 1995. The mean age was 41.4 +/- 14.4 years (range 14-85). There were 160 males (53%) and 142 females (47%). One hundred forty (46%) were type 1 and 162 (53.6%) were type 2. The mean duration of diabetes was 9.4 +/- 5.4 years and the mean Hemoglobin Alc (HbAlc) was 10.4 + 2.2%. On the day of the examination the mean fasting blood glucose (FBG) and random blood glucose (RBG) were 195.5 +/- 79.9 mg/dl and 273.1 +/- 114.5 mg/dl respectively. The mean serum total cholesterol, triglycerides, LDL, VLDL and GDL were 166.5 +/- 45.5 mg/dl, 129.9 +/- 92.4 mg/dl, 94.5 +/- 36.4 mg/dl, 24.4 +/- 15.1 mg/dl and 44.3 +/- 11.5 mg/dl respectively. The overall prevalence of retinopathy was 37.8% out of which 108 patients (36.1%) had background retinopathy and 5 patients (1.7%) had proliferative retinopathy. The retina could not be visualized in three patients because of dense cataract. Retinopathy correlated positively with age, duration of diabetes and blood pressure respectively, however no significant correlation was seen with mean total HgAlc and serum lipids. Prevalence of retinopathy was comparable in type 1 and type 2 (p > 0.05). The prevalence of retinopathy in our patients relative to the duration of diabetes mellitus is high. Therefore, improving facilities for the diagnosis and treatment of retinopathy is recommended.

Systematic inspection of insulin injection sites for local complications related to incorrect injection technique.

We systematically inspected insulin injection sites in 100 insulin-requiring patients attending the Diabetic Clinic of the Tikur Anbassa Hospital (TAH) in order to identify local complications related to incorrect injection technique: local complications were found in 53 cases: skin hyperpigmentation and/or indurations in 30 patients; and fat atrophy or hypertrophy in 31 patients. Illiteracy was significantly more common among those with local complications (18/53 versus 6/47, chi 2 5.03, P < 0.05). Mean fasting blood glucose on the day of the inspection was significantly higher (14.9 + 6.3 mmol/l versus 10.5 + 6.1 mmol/l, P < 0.001) and a fasting blood glucose > 10 mmol/l more common (41/53 versus 20/47, chi 2 14.1, P < 0.0005) in those with than in those without local complications. There was no significant difference between the two groups in mean duration of diabetes (6.9 + 5.6 years versus 6.6 + 5.8 years), frequency of hypoglycaemic episodes (12/53 versus 5/47, chi 2 1.76, P > 0.05) or mean daily insulin dose (44 + 18 units versus 44 + 22 units per day). Therefore, we concluded that local complications resulting from incorrect injection technique, a common finding in the group of patients studied, may be common among insulin requiring diabetic patients in general. Incorrect insulin injection causes local complications and disfigurement which may compromise compliance. Furthermore, insulin absorption tends to be erratic from intradermal and fat hypertrophy sites thus interfering with effective diabetic control. Insulin injection sites should be inspected routinely to detect and correct family technique promptly.