Abnormal Glucose Tolerance in Transfusion-Dependent β-Thalassemic Patients

Blood-transfused patients with β-thalassemia were recruited between January and May 2000 from the thalassemia clinics of National Taiwan University Hospital and Tao-Yuan General Hospital, Department of Health, which are the main hospitals in Taiwan that care for thalassemic patients. Those eligible for study were ≥2 years of age and were receiving frequent transfusions (15 ml packed erythrocytes per kg body wt given every 2–3 weeks to keep their hemoglobin level at a minimum of 10 g/dl before each transfusion). Taken at least 5 days each week, the dose of iron-chelating agent deferoxamine prescribed was 30–50 mg per kg body wt per day. The drug was given by subcutaneous infusion overnight for 8–12 h. There were 89 patients (demographic data are shown in Table 1) who met the criteria and consented to participate (i.e., they agreed to answer the questionnaire). Of the 89 patients, 14 were diagnosed with diabetes before the present study. In the remaining 75 patients, 68 were tested for blood glucose. All patients were interviewed and filled out standardized questionnaires during the first visit of the study, and their medical histories were obtained from the hospital files. This information included the following: age, sex, height, body weight, pubertal status according to Tanner’s classification, age at the first blood transfusion, duration of blood transfusion, age at the start of iron-chelation therapy, duration of iron-chelation therapy, compliance with iron-chelation therapy, serum ferritin level within 6 months, existence of hepatitis B surface antigen, existence of hepatitis C antibody, previous splenectomy, family history of diabetes, and previous bone marrow transplantation. The compliance with deferoxamine chelation therapy was assessed as good (≥5 deferoxamine infusions per week) in 59 subjects and poor (<4 infusions per week) in 23. A complete family history of diabetes noted in first-degree relatives and siblings was collected.

In this study, we used the World Health Organization’s definition of diabetes (13), which holds that subjects with regular treatment of diabetes are considered diabetic. If patients were not diabetic, an oral glucose tolerance test (OGTT) was performed in the morning after an overnight fast, as recommended by the National Diabetes Data Group (14). Glucose was ingested in a dose of 1.75 g/kg up to a maximum of 75 g, and blood samples were obtained at 0 and 120 min for measurement of plasma glucose. Patients ≥10 years of age were evaluated by a glucose tolerance test, whereas patients <10 years of age were evaluated by either a glucose tolerance test or a fasting glucose test, with the choice of test depending on the patients and their parents (15). An abnormal fasting glucose test result was followed by the administration of a glucose tolerance test (16).

Differences between patients with and without abnormal glucose tolerance were tested with the independent t test, χ² test, and Fisher’s exact test to identify the potential risk factors. In addition, multivariate analysis was performed using the stepwise logistic regression analysis to select the independent risk factors that best predicted abnormal glucose tolerance. A two-tailed P value of <0.05 was considered to be statistically significant.

During the study period, 89 patients (41 male and 48 female subjects) underwent regular blood transfusion in the thalassemia clinics of National Taiwan University Hospital and Tao-Yuan General Hospital, Department of Health. The age distribution of the patients is shown in Fig. 1. Although prenatal diagnosis for thalassemia had been initiated in 1993 in Taiwan, there were nonetheless 10 patients with thalassemia in this study who were born after 1994. These cases probably resulted from a new genetic mutation and poor coverage at the start of the prenatal screening program. There were also three “thalassemia intermedia” patients who were 7, 14, and 18 years of age at the start of blood transfusion. Of 82 patients, 59 (72.0%) had normal glucose tolerance, 7 (8.5%) had impaired glucose tolerance (IGT), and 16 (19.5%) had diabetes; 14 of the 16 diabetic subjects had been diagnosed previously. The characteristics of these different groups are shown in Table 2.

All seven patients with IGT were diagnosed on the basis of OGTT results. Mean age at diagnosis was 16.6 ± 4.9 years, mean age at the start of blood transfusion was 9.3 months (range 2–14 months), and mean age at the start of iron-chelation therapy was 9.6 years (4–24 years). The compliance was good in only two patients (28.6%). The serum ferritin concentration at the time of diagnosis ranged from 2,000 to 9,549 μg/l (mean 5,827.6). Three patients (42.9%) had hepatitis C infection, and none had hepatitis B infection. Splenectomy was performed in two patients (28.6%). No patients received bone marrow transplantation. Two patients (28.6%) had a family history of diabetes.

Of the 16 diabetic patients, 2 were diagnosed during the study and 14 were diagnosed previously. Mean age at the diagnosis was 17.4 years (range 7–24). Data about the manifestation of diabetes, available in 15 of 16 patients, indicated that diabetic ketoacidosis was the initial manifestation in 5 patients (31.1%) and asymptomatic hyperglycemia was the initial diagnosis in 11 patients (68.9%). Mean age at the start of blood transfusion was 5.75 months (1,2,3,4,5,6,7,8,9,10,11,12), and it was 9.8 years (4–19) at start of iron-chelation therapy. The compliance was good in nine patients (56.3%). The serum ferritin concentration, available in 13 patients at the time of diagnosis, ranged from 1,180 to 13,500 μg/l (mean 5,299.6). Of the 15 patients for whom data on viral hepatitis were available, none had hepatitis B, and 9 (56.3%) had hepatitis C infection. Eight patients (50%) underwent splenectomy and three (18.8%) received bone marrow transplantation. Four patients (25%) had a positive family history of diabetes.

Blood glucose levels of patients newly diagnosed with IGT and diabetes are shown in Table 3. In the patients with IGT, fasting glucose levels of two patients (28.6%) were <110 mg/dl; in patients with diabetes, the fasting glucose level of one patient (50%) was <126 mg/dl. We therefore consider the American Diabetes Association (ADA) criterion (17) to be unsuitable for diagnosis of IGT or diabetes in this rare disease group (18).

After combining the patients with IGT and those with diabetes, we determined which patients had abnormal glucose tolerance levels. We then grouped the patients into those having normal and those having abnormal glucose tolerance. The risk factors associated with abnormal glucose tolerance were assessed (Table 4). Univariate analyses showed the following variables to be significantly associated with abnormal glucose tolerance: age of patient, height, body weight, age at the first transfusion, duration of blood transfusion, age at the start of iron-chelation therapy, compliance with iron-chelation therapy, serum ferritin concentration, hepatitis C infection, and previous splenectomy. These variables were included in the stepwise logistic regression. After adjustment for confounding factors, serum ferritin concentration (odds ratio [OR] 1.0005, P = 0.006) and hepatitis C infection (OR 3.8343, P = 0.0491) were identified as independent risk factors for abnormal glucose tolerance (Table 5).

Thalassemia is a major health concern in Taiwan; the carrier rate of β-thalassemia is at least 1.1% (19,20). The Department of Health of Taiwan has had a national thalassemia-major screening program since 1993 (21). Nevertheless, 10 children with thalassemia major were born after 1994. Consequently, the efficacy of prenatal screening programs in Taiwan is under scrutiny.

This study is the first to report the relationship between glucose tolerance and blood-transfused β-thalassemia in Taiwan. Adequate iron-chelation therapy can prevent complications, including diabetes (22,23,24,25). Serum ferritin concentration (OR 1.0005, P = 0.006), a marker for hepatic iron concentration, was found to be an independent risk factor in this study. Serum ferritin has been considered to be a component of insulin resistance syndrome in epidemiological studies (26). Our results in this rare disease group lend further support to this viewpoint.

Studies have indicated that when the serum ferritin level is <2,500 μg/l, the incidence of diabetes is reduced (15,22,23). By contrast, we found that three patients with diabetes and one patient with IGT had serum ferritin levels of 1,180, 1,475, 1,580, and 2,000 μg/l, respectively. Although serum ferritin plays an important role in the genesis of insulin resistance, we believe that the development of diabetes might be complicated by other conditions.

Diabetes has been observed to develop shortly after an episode of acute viral hepatitis in thalassemic patients (11), and chronic hepatitis C infection was considered to have a diabetogenic effect (27). None of our patients had acute hepatitis, but in our patients with chronic infection, hepatitis C was found to be an independent risk factor for abnormal glucose tolerance (OR 3.8343, P = 0.0491). Because iron-induced hepatic damage is exacerbated by infection with hepatitis C virus (28,29), hepatic dysfunction may be the most important cause for the development of insulin resistance (12,15,21,30) and abnormal glucose tolerance. Therefore, the interplay of iron burden and hepatitis C may be the determining factor in the development of abnormal glucose tolerance. We believe that age and duration of blood transfusion merely determine the total amount of blood transfused. When that amount increases, the possibility of iron overload and infection with hepatitis C virus is increased.

It is still unclear whether diabetes in β-thalassemia major is related to genetic factors (6,10,12,15,32,33). We did not demonstrate that family history was a risk factor in our patient group.

Many drugs used for bone marrow transplantation may be diabetogenic, but our data showed that previous bone marrow transplantation (33,34,35) was not a risk factor in the development of abnormal glucose tolerance.

The mechanism of abnormal glucose homeostasis in patients with β-thalassemia major is still unknown but is attributed mainly to insulin deficiency resulting from the toxic effects of iron deposited in the pancreas (5,6,31) and from insulin resistance (12,15,28). Insulin resistance may come from iron deposition in both liver (where iron deposits may interfere with insulin’s ability to suppress hepatic glucose production) and muscle (where iron deposits may decrease glucose uptake because of muscle damage) (12). Persistent insulin resistance along with a progressive reduction in circulating insulin levels may lead to glucose intolerance and overt diabetes (36).

Early identification of thalassemic patients with IGT has decreased presentations with diabetic ketoacidosis, but these presentations remain as high as 31.1% in Taiwan, according to this study. As a result, physicians caring for patients with thalassemia major should be particularly alert to the possibility of diabetes. Because not all of the patients with thalassemia major could be correctly diagnosed by fasting glucose alone (18), we preferred to use OGTT rather than the guidelines of the ADA (17) for the diagnosis of abnormal glucose tolerance in thalassemic patients.

In conclusion, we found that serum ferritin concentration and hepatitis C infection are risk factors of abnormal glucose tolerance in patients with blood-transfused β-thalassemia. Aggressive iron-chelation therapy and prevention and treatment of hepatitis C infection are now the most important issues in managing glucose homeostasis in transfusion-dependent β-thalassemic patients in Taiwan.