Apolipoproteins and lipoproteins in children with type I diabetes: relation to glycosylated serum protein and HbA1

Serum levels of cholesterol (C), triglycerides (TG), lipoprotein-C and apolipoproteins (apo) A-I, A-II and B were measured in 30 children with type I diabetes mellitus (16 boys, 14 girls, aged 11-14 years) and in 26 healthy controls (15 boys, 11 girls, aged 10-13 years). For 19 diabetics controls matched for age, sex and relative body weight were selected. The diabetic patients were considered to be in fair metabolic control according to HbA1 levels and glycosylated serum protein concentrations. Mean serum apo A-I, A-II and B, C, TG, low density lipoprotein cholesterol (LDL-C) and high density lipoprotein cholesterol (HDL-C) did not differ significantly between diabetic nondiabetic children. Very low density lipoprotein cholesterol (VLDL-C) was significantly higher in diabetic children than in controls. Serum C and LDL-C levels showed close univariate linear correlations with glycosylated serum protein (LDL-C: r = 0.53, p less than 0.01, C: r = 0.58, p less than 0.01) in diabetics. The ratio LDL/HDL-C was significantly correlated to HbA1 levels (r = 0.47, p less than 0.01). By canonical and multiple linear correlation analysis significant relations of a selected set of variables concerning the control and therapy of diabetes (serum glucose, HbA1, glycosylated serum protein, insulin dose) with a set of lipoprotein variables (C, TG, VLDL-C, HDL-C, LDL-C, apo A-I, A-II, B) could be demonstrated. From these data we conclude that significant relations between atherogenic serum lipids and lipoproteins (C, LDL-C) and the degree of metabolic control exist in diabetic children, even in the absence of marked dyslipoproteinemia.(ABSTRACT TRUNCATED AT 250 WORDS)     

Apolipoproteins and Lipoproteins in Children with Type I Diabetes: Relation to Glycosylated Serum Protein and HbA1

ABSTRACT. Serum levels of cholesterol (C), triglycerides (TG), lipoprotein-C and apolipoproteins (apo) A-I, A-II and B were measured in 30 children with type I diabetes mellitus (16 boys, 14 girls, aged 11–14 years) and in 26 healthy controls (15 boys, 11 girls, aged 10–13 years). For 19 diabetics controls matched for age, sex and relative body weight were selected. The diabetic patients were considered to be in fair metabolic control according to HbA₁ levels and glycosylated serum protein concentrations. Mean serum apo A-I, A-II and B, C, TG, low density lipoprotein cholesterol (LDL-C) and high density lipoprotein cholesterol (HDL-C) did not differ significantly between diabetic and nondiabetic children. Very low density lipoprotein cholesterol (VLDL-C) was significantly higher in diabetic children than in controls. Serum C and LDL-C levels showed close univariate linear correlations with glycosylated serum protein (LDL-C: r =0.53, p <0.01, C: r =0.58, p <0.01) in diabetics. The ratio LDL/HDL-C was significantly correlated to HbA₁ levels ( r =0.47, p <0.01). By canonical and multiple linear correlation analysis significant relations of a selected set of variables concerning the control and therapy of diabetes (serum glucose, HbA₁, glycosylated serum protein, insulin dose) with a set of lipoprotein variables (C, TG, VLDL-C, HDL-C, LDL-C, apo A-I, A-II, B) could be demonstrated. From these data we conclude that significant relations between atherogenic serum lipids and lipoproteins (C, LDL-C) and the degree of metabolic control exist in diabetic children, even in the absence of marked dyslipoproteinemia. The close relation of LDL-C and total C with glycosylated serum protein in the diabetics might be due to glycosylation of LDL .     

Increased high density lipoproteins in diabetic children

Serum lipoprotein lipid and apolipoprotein concentrations were determined in 27 diabetic children (5–18 years old) and 13 matched healthy controls. The serum cholesterol concentrations in the diabetics were slightly higher than in the controls (P<0.05) due to a significantly higher level of the high density lipoprotein cholesterol (P<0.01). Also the serum concentration of apolipoprotein A-I, the major protein constituent of the high density lipoprotein fraction, was higher in the diabetic children (P=0.05). There were no significant differences between the groups with regard to the serum triglyceride concentrations or the apolipoprotein C-II and C-III concentrations. Neither the lipoprotein lipid nor the apolipoprotein levels were significantly correlated with variables related to the degree of regulation of the diabetic disease. No obvious explantation, based on the present data, can be given for the increased high density lipoprotein cholesterol concentrations in insulin-treated diabetics in comparison with the healthy children. It is possible, however, that the increased high density lipoprotein cholesterol concentration may be caused by an increased level of insulin in the circulation of insulin-treated diabetic children.     

Reduced serum proteins in diabetic children on a twice-daily insulin schedule

The concentrations of selected proteins of transport and nutrition were investigated in 27 diabetic children and 13 healthy controls by an immunodiffusion technique. The diabetic children had significantly lower mean pre-albumin (p less than 0.001), albumin (p less than 0.01) and orosomucoid levels (p less than 0.05) than the healthy controls. No correlation was observed between age or sex and the blood concentrations of the specific proteins analyzed in this series. Haptoglobin and hemopexin showed positive correlations with serum triglycerides (both p less than 0.01) and slight positive correlations with some of the variables of carbohydrate control. The reduced levels of prealbumin and albumin were not correlated to diabetic control as measured by HbA1, fasting plasma glucose or urinary glucose excretion. The non-physiological distribution site and the abnormal temporal pattern of insulin offered to diabetic children might be the reason for the protein abnormalities found in this study. This is seemingly the first time reduced serum levels of proteins have been reported in diabetic children.     

SERUM LIPIDS AND LIPOPROTEINS IN 157 INSULIN DEPENDENT DIABETIC CHILDREN AND ADOLESCENTS IN RELATION TO METABOLIC REGULATION, OBESITY AND GENETIC HYPERLIPOPROTEINEMIA

ABSTRACT. Serum lipids and lipoproteins were measured in 157 insulin dependent diabetic children and adolescents (IDDM) and in 350 healthy reference individuals. Serum triglyceride values were lower and total cholesterol and high density lipoprotein cholesterol higher in IDDM. Metabolic regulation reflected by glucosuria, postprandial blood glucose, number of hypoglycemic episodes and hemoglobin AIC all correlated strongly with serum triglyceride and very low density lipoprotein cholesterol but not with serum low density lipoprotein cholesterol or high density lipoprotein cholesterol. Serum lipids and lipoproteins did not correlate with obesity. Three children had genetic hyperlipoproteinemia.
In IDDM measurement of serum lipids and lipoproteins can thus be used to further assess metabolic regulation. Measurement of serum lipids and lipoproteins seems warranted for future evaluation of the risk of cardiovascular disease in IDDM.     

HYPOMAGNESEMIA IN DIABETIC CHILDREN

ABSTRACT. Serum concentrations of iron, copper, zinc and magnesium were determined in 27 Swedish children with well controlled diabetes without longterm diabetic complications. Thirteen of the diabetic children had age- and sex-matched healthy controls. Significantly lower serum magnesium concentrations were found in the diabetic children than in the matched healthy controls ( p <0.01). The levels of iron, copper and zinc did not differ in the diabetic children from those in the controls. A negative correlation between serum magnesium level and duration of diabetes was found ( p <0.05). This is seemingly the first report of hypomagnesemia in diabetic children.     

High density lipoprotein cholesterol and total cholesterol in children with asthma and allergic rhinitis

Serum high density lipoprotein cholesterol and total cholesterol were studied in 15 children with allergic rhinitis, 45 asymptomatic asthmatic children and 16 children with acute asthmatic attacks. The latter were also studied in their asymptomatic phase. Two control groups of children with minor anatomical defects and those with acute pneumonia with respiratory distress were also studied. The serum concentrations of high density lipoprotein cholesterol were significantly higher (p less than 0.001) and those of total cholesterol lower (p less than 0.001) in children with respiratory allergy when compared to controls. The concentration of high density lipoprotein cholesterol increased and that of total cholesterol decreased during acute asthma.     

脓毒症并多器官功能障碍综合征的血脂分析与预后关系的探讨

目的探讨血脂在脓毒症并多器官功能障碍综合征(multipleorgandysfunctionsyn-drom,MODS)患儿血中的变化及与预后的关系。方法选择符合诊断标准的34例脓毒症并MODS的患儿(其中最后存活22例,死亡12例)作为脓毒症组,检测确定诊断后第2天晨起空腹血的血脂水平,并选择同一时期同一年龄组健康体检者15例,测血脂水平做为对照,进行前瞻性研究。结果存活者与死亡者的血清胆固醇(总胆固醇、低密度脂蛋白胆固醇、高密度脂蛋白胆固醇)相比差异有显著性(P<0·01),死亡者有较低浓度的血胆固醇水平;脓毒症组与对照组的血清胆固醇(总胆固醇、低密度脂蛋白胆固醇、高密度脂蛋白胆固醇)相比差异有显著性(P<0·01),脓毒症组有较低浓度的血胆固醇水平,而血甘油三酯无显著性差异。结论脓毒症并MODS患儿的血胆固醇(总胆固醇、低密度脂蛋白胆固醇、高密度脂蛋白胆固醇)水平与预后有关,血胆固醇(总胆固醇、低密度脂蛋白胆固醇、高密度脂蛋白胆固醇)越低,预后越差。     

Serum lipids and lipoproteins in 157 insulin dependent diabetic children and adolescents in relation to metabolic regulation, obesity and genetic hyperlipoproteinemia

Serum lipids and lipoproteins were measured in 157 insulin dependent diabetic children and adolescents (IDDM) and in 350 healthy reference individuals. Serum triglyceride values were lower and total cholesterol and high density lipoprotein cholesterol higher in IDDM. Metabolic regulation reflected by glucosuria, postprandial blood glucose, number of hypoglycemic episodes and hemoglobin A1c all correlated strongly with serum triglyceride and very low density lipoprotein cholesterol. Serum lipids and lipoproteins did not correlate with obesity. Three children had genetic hyperlipoproteinemia. In IDDM measurement of serum lipids and lipoproteins can thus be used to further assess metabolic regulation. Measurement of serum lipids and lipoproteins seems warranted for future evaluation of the risk of cardiovascular disease in IDDM.     

Elevated serum selenium in diabetic children

Twenty-seven diabetic children, 16 girls and 11 boys, 5-18 years of age, with a duration of the disease ranging from 2-15 years, comprised the study group. Thirteen children with a similar age and sex distribution, living in the same area served as healthy controls. All 40 children had a normal growth pattern. The mean serum selenium concentration in the diabetic children, determined by neutron activation analysis, was 7.4 +/- 0.8 micrograms/100 ml (mean +/- SD) and in the healthy controls 6.5 +/- 0.8 micrograms/100 ml. The difference between the two groups was statistically highly significant (p less than 0.01). Boys and girls in both groups had nearly identical mean serum selenium levels and no correlation was observed between the selenium concentrations and either the age, weight or height of the children or the indicators of diabetic control. The selenium status in diabetic children has not been reported previously. The possibility of elevated serum selenium in diabetic children in response to altered lipid metabolism is discussed.     

Euthyroid sick syndrome in type I diabetes mellitus in children and adolescents

We studied concentrations of thyroid hormones (T3, T4, FT4, rT3, TBG and TSH) in 62 type I diabetic children and adolescents. The patients were classified into group A (n = 27, good control, HbA1c less than 10%), group B (n = 19, poor control, HbA1c greater than 10%) and group C (n = 16, diabetic ketoacidosis, pH less than 7.1 and HCO3 less than 15 mmol/L. All patients were treated with two daily injections of purified monocomponent insulins. Thirty healthy subjects of the same age served as control group. Patients in group B and C had significantly lower T3 and higher rT3 levels (p less than 0.001) compared to the matched controls (1.5 vs 2.2; 0.9 vs 2.2; 0.58 vs 0.3 and 0.6 vs 0.3 nmol/L). Serum TBG levels were significantly lower (p less than 0.01) in the group A (19.5 +/- 4.3 mg/L), group B (20.3 +/- 3.3) and group C (18.0 +/- 3.4) compared with control group (24.2 +/- 3.1). There was significantly negative correlation between T3 and HbA1c in group B (r = 0.546; p less than 0.02). The results of this study confirm that euthyroid sick syndrome does exist in type I diabetic children and adolescents with poor metabolic control and ketoacidosis. The inverse relationship between T3 and HbA1c percentage (low T3 and high HbA1c) points to the poor diabetic control.     

Osteoporosis in cystic fibrosis

To determine if osteoporosis is prevalent among patients with cystic fibrosis, we compared the vertebral bone density measured by quantitative computed tomography in 57 such patients (29 male, 28 female, aged 3 to 21 years) with those of an age-, race-, and sex-matched control group of 57 healthy subjects. Patients with cystic fibrosis had significantly lower bone density (10% lower, p less than 0.001) than in controls. The decrease in bone density in patients with cystic fibrosis was unrelated to age. Shwachman clinical evaluation scores (based on case history, pulmonary physical findings, growth, and x-ray findings) correlated positively with age-standardized bone density values (p less than 0.01). Male patients had substantially lower bone density than did female patients (p less than 0.02), but bone density differences related to gender were not significant when effects of disease severity were controlled for. Decreased bone density was more common in patients with poor nutritional status as determined by anthropometric measurements (p less than 0.05). We conclude that osteoporosis is a frequent complication in children with cystic fibrosis regardless of their age and is more prevalent in patients with greater disease severity.     

Parental history of cardiovascular disease as an indication for screening for lipoprotein abnormalities in children

We studied the relationship between parental history of cardiovascular disease and risk for adverse lipid and lipoprotein levels in a total community study of 3313 children (ages 4 to 17 years, 63% white, 37% black). Older white children (11 to 17 years) with a parental history of heart attack or diabetes were 4.3 and 5.6 times, respectively, more likely to have high levels (greater than or equal to 95th percentile) of serum total cholesterol than those without such a history (all p less than 0.05). White children with a parental history of heart attack or diabetes were twice as likely to have an elevated (greater than or equal to 95th percentile) low-density lipoprotein cholesterol (LDL-C) level than those without such a history (both p less than 0.05). In contrast, parental history of cardiovascular disease did not predict elevated levels of total cholesterol or LDL-C in black children. However, older black children with a parental history of heart attack, hypertension, or diabetes were 4 1/2 to 5 times more likely to have low levels (less than or equal to 5th percentile) of high-density lipoprotein cholesterol than those without such a history (all p less than 0.05). Only 40% of white children and 21% of black children with elevated LDL-C levels had a parental history of vascular disease. These findings raise questions about the current practice of screening only children with a family history of cardiovascular disease to identify those with elevated total cholesterol and LDL-C levels.     

High density lipoprotein-cholesterol changes in children with high cholesterol levels at birth

The predictive value of serum lipoprotein concentrations at birth for the same parameters later in life is under debate. A group of 20 children displaying high total cholesterol (TC) levels at birth (group 2) were compared at age 4 years with 18 control children who had presented a normal lipoprotein profile at birth (group 1). There was a significant correlation between TC, low density lipoprotein-cholesterol, high density lipoprotein (HDL)-cholesterol, and apolipoprotein (Apo) A-I levels at age 4 years and at birth. The increases in TC and HDL-cholesterol levels from birth to age 4 years were significantly lower (P < 0.05, P < 0.01, respectively) in group 2 than in the control group and inversely correlated with the concentrations of these parameters at birth. The increases in HDL-cholesterol and Apo A-I levels were higher in males while those of triacylglycerol and Apo B were higher in females (P < 0.05). However, the increases in TC and HDL-cholesterol were higher in controls (P< 0.05). Diets of children of both groups were similar regarding the energy contribution of saturated, monounsaturated and polyunsaturated fatty acids, although children from group 2 ate less fish and omega-3 fatty acids (P < 0.05). CONCLUSION: the present data suggest for the first time that when high density lipoprotein-cholesterol levels are high at birth, those levels increase less during the first four years of life. Moreover, low density lipoprotein-cholesterol increased about five times as much as high density lipoprotein-cholesterol did in controls and about 15 times as much as in the children with high cholesterol at birth.     

Apolipoprotein E phenotypes and plasma lipids in diabetic children and adolescents

Apolipoprotein E (apoE) polymorphism is a genetic determinant of serum lipoprotein levels and coronary heart disease risk. ApoE appears in three major isoforms E2, E3 and E4, coded by corresponding alleles 2, 3 and 4. These give six different phenotypes. Patients with insulin dependent diabetes (IDDM) have been reported to have increased incidence of E2/2 homozygosity. We studied the frequencies of apoE phenotypes and their association with plasma lipids in 201 diabetic children, aged 2–17 years, and in 216 healthy controls with the same age range. Phenotyping was performed directly from plasma by iso-electric focusing and immunoblotting. Plasma total and high density lipoprotein (HDL) cholesterol (C) and triglycerides were determined by routine laboratory methods. Apolipoprotein A1 (apoA1) and B (apoB) were measured by turbidometry. There were no differences in apoE phenotype or allele distributions between the diabetic and control subjects. The frequencies of 2, 3, and 4 in the diabetic and control children were 0.08 versus 0.07, 0.73 versus 0.72 and 0.19 versus 0.21. The difference in apoE2/2 frequencies (2.0 in diabetic and 0.5% in normal children) was not statistically significant. In the diabetic children, there was a distinct relation between apoE phenotype and plasma lipids; presence of apoE2 was associated with the lowest and that of apoE4 with the highest concentrations of total and low density lipoprotein (LDL) C, and apoB. Ratios of HDL-C/LDL-C and apoA1/apoB showed on opposite trend. The influence of apoE polymorphism on plasma lipids was less clear in the controls. ApoE polymorphism did not relate to body mass index, glycohaemoglobin A1c or daily insulin dosage. As compared to controls, diabetic children had significantly lower concentrations of total and LDL-C and apoB as well as higher ratios of HDL-C/LDL-C and apoA1/apoB. In conclusion, there was no difference in distribution of apoE phenotypes or alleles between diabetic and non-diabetic children. ApoE polymorphism proved to be an important genetic determinant of plasma lipid levels in patients with IDDM.     

Metabolic evaluation of obese and nonobese siblings

OBJECTIVE: To test the hypotheses that obese adolescents have a lower resting metabolic rate and less aerobic endurance than their nonobese siblings. DESIGN: Case-referent study of obese and nonobese siblings from the same kindred. SETTING: Tertiary referral center. PARTICIPANTS: Telephone screening of community volunteers resulted in a consecutive sample of 16 kindreds. Obese and nonobese siblings were similar in age, height, and pubertal status. Significantly more female subjects were in the obese group (p less than 0.01). MEASUREMENTS AND MAIN RESULTS: Body composition studies revealed that the obese siblings had higher body fat (p less than 0.001) but that fat-free mass was similar to that of the lean siblings. Resting metabolic rates determined by indirect calorimetry for the obese and nonobese pairs did not differ. Although the obese siblings appeared less fit when maximal oxygen consumption was measured in relation to total weight, maximal oxygen consumption did not differ when values were standardized for fat-free mass. CONCLUSIONS: The obese adolescents did not have a reduced resting metabolic rate. As in adults, the relationship between resting metabolic rate and fat-free mass was similar for obese and nonobese children and adolescents. Any decreased sport participation by the obese siblings was not due to inherent reductions in aerobic capacity.     

Type 1 diabetic children have abnormal lipid profiles during pubertal years

Aim/Hypothesis: To study the prevalence of hypercholesterolemia, hypertriglyceridemia and the relationship between metabolic control, pubertal status and plasma lipoprotein levels in children with diabetes mellitus.
Subjects and methods: A cross‐sectional study was conducted on 126 subjects with type I diabetes followed at our institution. There were 57 boys and 69 girls (mean age: 13.4±3.4 yr; mean duration of diabetes: 7.3±2.1 yr), on whom fasting lipoprotein levels and pubertal status were determined. Mean glycated hemoglobin (HbA1c) of the preceeding year was used in the analysis. Cholesterol (CT) and triglyceride (TG) levels were transformed into standard deviations (SD) using age dependent normal values.
Results: 1) CT levels of DM children (mean level: +0.9±1.2 SD) are higher for both sexes and at each age. Sixteen percent of the cases had CT level ≥2 SD. Within the range of the HbA1c observed (9.1±1.2%), CT levels are not correlated with the degree of metabolic control. In contrast to non‐diabetic children, CT levels of the diabetic children did not vary throughout pubertal stages. CT levels correlated highly with apolipoprotein B (r=0.79; p<0.00001 and r²=82%, in univariate and multivariate analysis, respectively. 2) Plasma TG levels are comparable in the diabetic children (mean level: ?0.11±0.9 SD) and non‐diabetic children. Only 5% of the diabetic children have a TG level ≥2 SD. The TG levels are significantly, but weakly, positively correlated with duration of diabetes and the degree of metabolic control (r²=12% and 16%, respectively, p<0.0001 for both).
Conclusions: Plasma CT levels of type I diabetic children are increased in comparison to non‐diabetic children and do not follow the usual decreasing pattern during puberty.     

Does VLDL-LDL-cholesterol in cord serum predict future level of lipoproteins?

Lipoproteins were measured in 618 healthy, full-term newborns. Seventy-four with a VLDL-LDL-cholesterol above 1.3 mmol/l (50 mg/dl) at birth and 25 randomly chosen controls with VLDL-LDL-cholesterol 1.3 mmol/l or below at birth were followed up at age 2. Seventy-two (52 in the high VLDL-LDL-cholesterol group and 20 in the low VLDL-LDL-cholesterol group) were followed up at age 13. At age 2 mean total cholesterol was 5.48 mmol/l (SEM 0.10) in the children with a high VLDL-LDL-cholesterol at birth, compared to 4.69 mmol/l (SEM 0.17) in the children with a low VLDL-LDL-cholesterol at birth (p less than 0.001). A difference was still present at age 13 (4.74 mmol/l; SEM 0.11 versus 4.20 mmol/l; SEM 0.14; p less than 0.01). At age 13 apolipoprotein B was 0.74 g/l (SEM 0.02) in the children with a high VLDL-LDL-cholesterol at birth, compared to 0.65 g/l (SEM 0.02) in the children with a low VLDL-LDL-cholesterol at birth (p less than 0.01). Children with high VLDL-LDL-cholesterol at birth might be more liable to high lipoprotein serum levels later in life.     

Assessment of magnesium status in newly diagnosed diabetic children: measurement of erythrocyte magnesium level and magnesium tolerance testing

The aim of this study was to investigate the relationship between serum, erythrocyte and urine magnesium levels and retained magnesium percentage in newly diagnosed diabetic children. In a cross-sectional study, 34 children with insulin dependent diabetes mellitus (IDDM) and 21 healthy age- and sex-matched control subjects were screened for their serum, erythrocyte, and urine magnesium levels. Magnesium tolerance test was performed on diabetic and control subjects. Serum and erythrocyte magnesium levels in diabetic children were significantly lower than in healthy controls (plasma magnesium, p<0.05; erythrocyte magnesium, p<0.001); however, serum magnesium level was in normal range in diabetics and controls. Erythrocyte magnesium levels in diabetic children showed an inverse correlation with percentage of retained magnesium load (r=-0.44, p<0.01). Urine magnesium excretion in diabetic children (7.12 +/- 2.18 mmol/g creatinine/24-hr) was significantly higher than in healthy controls (4.0 +/- 1.35 mmol/g creatinine/24-hr) (p<0.001). There was a negative correlation between erythrocyte magnesium (2.07 +/- 0.62 mmol/L) and urine magnesium (7.12 +/- 2.18 mmol/g creatinine/24-hr) (r=-0.68 p<0.01) in diabetic children. Magnesium tolerance test showed that percentage of retained magnesium in diabetic children (66 +/- 26%) was significantly higher than in controls (16 +/- 7%) (p<0.001). This study is the first study to simultaneously investigate serum, erythrocyte and urine magnesium levels and magnesium tolerance test in newly diagnosed diabetic children. In conclusion, erythrocyte magnesium levels decrease earlier than serum magnesium in diabetic children. The follow-up parameters in diabetics may include the policy of monitoring magnesium status. Erythrocyte magnesium measurement is preferred to serum magnesium. Magnesium tolerance test is a reliable and sensitive method, which may be used as an alternative to erythrocyte magnesium measurement or in combination with it in hospitalized diabetic children.     

Treatment of children with homozygous familial hypercholesterolemia: Safety and efficacy of low-density lipoprotein apheresis

We evaluated the safety and efficacy of dextran sulfate low-density lipoprotein (LDL) apheresis in the treatment of three children (aged 6, 7, and 10 years) with severe familial homozygous hypercholesterolemia and undetectable LDL receptor activity. A total of 35 double plasma volume procedures were performed. The ranges of the mean decreases of the three patients in plasma lipid concentrations after LDL apheresis (p less than 0.0001) were as follows: total cholesterol, 76% to 79%; LDL-cholesterol, 78% to 81%; very low density lipoprotein cholesterol, 69% to 75%; high-density lipoprotein cholesterol, 27% to 40%; and triglycerides, 34% to 68%. There were statistically significant but clinically and biologically irrelevant changes in hematologic indexes, serum chemistry values, immunoglobulin levels, complement activity, and plasma concentrations of fat-soluble vitamins. Simple correlation analysis of the variables affecting total cholesterol removal showed significant correlation coefficients (r values) for preapheresis total cholesterol values (r = 0.70; p less than 0.01) and preapheresis LDL-cholesterol values (r = 0.61; p less than 0.01). A multiple regression model explained 82% of the variance based on the preapheresis cholesterol concentration, volume of whole blood processed, and the serum albumin concentration. Side effects of the LDL-apheresis treatments were rare and included abdominal cramping and urticaria. Two procedures were aborted because of intravenous access problems in the younger children. This study confirms that LDL apheresis using a dextran sulfate affinity column is efficacious in rapidly lowering total and LDL-cholesterol concentrations. Furthermore, the procedure is safe and well tolerated by children as young as 6 years of age. This treatment may prevent the progression of atherosclerosis in children with homozygous familial hypercholesterolemia and may therefore avert early death.