ENDOCRINOLOGY

T. Ohta á S. Nishiyama á T. Nakamura á K. Saku á K. K. Maung á I. Matsuda

Predominance of large low density lipoprotein particlesand lower fractional esteri®cation rate of cholesterolin high density lipoprotein in childrenwith insulin-dependent diabetes mellitus

Received: 27 March 1997 /Accepted in revised form: 18 September 1997

Abstract Coronary artery disease (CAD) is a major cause of death in patients withinsulin-dependent diabetes mellitus. Qualitative changes in low density lipoprotein (LDL)and high density lipoprotein (HDL) are thought to be important for evaluating the riskfor CAD. In the present study, we evaluated LDL particle size (LDL-size) by 2%±16%gradient gel electrophoresis, along with conventional lipids and apolipoproteins, in 23children with IDDM (10 males and 13 females) and 27 nondiabetic controls (12 males and15 females). The fractional and molar esteri®cation rates (FER and MER) of cholesterolin plasma and HDL were also determined. Plasma levels of triglyceride were signi®cantlylower in diabetic children than in controls. Plasma apoA-I and apoA-II levels in femalediabetic children were signi®cantly higher and lower than those in controls respectively.Plasma levels of HDL-cholesterol and the ratio of apoA-I to apoA-II were signi®cantlyhigher in diabetic children than in controls. Other lipid and apolipoprotein parameters indiabetic children were similar to those in controls. LDL-size in diabetic children wassigni®cantly greater than that in controls. FERHDL, which re¯ects the particle size dis-tribution of HDL, was signi®cantly lower in diabetic children than in controls, whichsuggests that diabetic children had larger HDL particles.

Conclusion The qualitative and quantitative changes in LDL and HDL in diabeticchildren are similar to those associated with a reduced risk for CAD. Intensive insulintherapy in children may help preventing coronary heart disease in adulthood.

Key words Children with IDDM á LDL-size á FERHDL á Coronary artery diseases

Abbreviations CAD coronary artery disease á IDDM insulin-dependent diabetes mellitus áNIDDM non-insulin-dependent diabetes mellitus á LDL low density lipoprotein á HDLhigh density lipoprotein á FERHDL fractional esteri®cation rate of cholesterol in HDL áMERHDL molar esteri®cation rate of cholesterol in HDL á FERplasma fractionalesteri®cation of cholesterol in plasma á MERplasma molar esteri®cation of cholesterol inplasma á LCAT lecithin cholesterol acyltransferase

Introduction

Plasma lipoproteins play an important role in the de-velopment of atherosclerosis [11, 22, 26, 32]. In patients

with non-insulin-dependent diabetes mellitus (NIDDM),an association of low density lipoprotein (LDL) andhigh density lipoprotein (HDL) with atheroscleroticcoronary artery disease (CAD) has been shown[4, 9, 28, 43]. CAD is also one of the major causes of

Eur J Pediatr (1998) 157: 276±281 Ó Springer-Verlag 1998

T. Ohta (&) á S. Nishiyama á T. NakamuraK. K. Maung á I. MatsudaDepartment of Paediatrics, Kumamoto University Schoolof Medicine, Honjo 1-1-1, Kumamoto City,Kumamoto 860, Japan,

e-mail: ohta@gpo.kumamoto-u.ac.jp,Tel.: 81-96-373-5191, Fax: 81-96-366-3471

K. SakuDepartment of Internal Medicine, Fukuoka University Schoolof Medicine, Fukuoka 814-01, Japan

death in patients with insulin-dependent diabetes mell-itus (IDDM) [20, 27, 30], but the association betweenlipoproteins and CAD is not yet clear. In our previousreport, we showed that the plasma levels of HDL andsubspecies of HDL in IDDM children are similar orhigher than those in a nondiabetic population conveyinga reduced risk for CAD. However, the qualitativechanges in HDL have not yet been studied [38].

With respect to qualitative changes in LDL andHDL, the particle size of these lipoproteins seems to beimportant. Elevated plasma levels of low density lipo-protein cholesterol (LDL-C) are associated with a in-creased risk for CAD [11, 26]. However, several studieshave shown that 30%±40% of CAD patients have nor-mal LDL-C levels [21, 24, 36]. Recent analyses of LDLsubclasses have indicated that smaller LDL particles arepredominant in a high proportion of CAD patients withnormal LDL-C levels [3, 10, 12, 13]. As with LDL, theparticle size of HDL also seems to be important forevaluating the risk for CAD. Some genetic syndromeswith HDL de®ciency are not associated with CAD, suchas Tangier disease and familial lecithin:cholesterolacyltransferase (LCAT) de®ciency without renal disease[37, 44]. Recently, Rader et al. [42] reported healthyasymptomatic subjects with very low HDL-C (<15mg/dl). These ®ndings are not consistent with a simpleinverse correlation between plasma HDL levels and therisk for CAD, but can be explained by a normal pro-portion of atherogenic and anti-atherogenic HDL frac-tions in these subjects [37, 42, 44]. They suggest that ananalysis of the particle size of both LDL and HDL isimportant for predicting the risk for CAD.

The fractional esteri®cation rate (FER) of cholesterolin very low density lipoprotein (VLDL) and LDL-de-pleted plasma (FERHDL), which re¯ects the reactivity ofHDL to lecithin:cholesterol acyltransferase (LCAT),shows a strong positive correlation with plasma levels ofHDL3b,c particles and a strong negative correlation withthe concentration of HDL2b and large LpA-I (HDLcontaining only apoA-I) [17, 18, 39]. Therefore,FERHDL can be used to evaluate the particle size dis-tribution and physicochemical characteristics of HDL inplasma samples [18]. In the present study, to evaluate thequalitative changes in LDL and HDL, we determinedthe LDL-size and FERHDL in children with IDDM.

Subjects and methods

Subjects

The clinical data for 23 Japanese children (10 boys and 13 girls)with IDDM are summarized in Table 1. The diagnosis of IDDMwas based on the criteria of the National Diabetes Data Group[34]. All subjects had been diagnosed 2±7 years earlier and wereclinically stable at the time of the study. Insulin (a combination oflong-acting and rapid human insulin) was given subcutaneouslyfour times a day. All patients were on adequate diets (caloric in-take: 50% from carbohydrate, 30% from fat and 20% from pro-tein). Twenty-seven non-diabetic healthy children (12 boys and 15girls) served as controls. They were outpatients of the Departmentof Paediatrics, Kumamoto University School of Medicine, who hadvisited the hospital for a routine check-up. The age distribution inthe control group was similar to that in the diabetic patients. Noneof the study subjects had renal failure or hepatic or thyroid ab-normalities and no child was taking any drug (except for insulin bythe diabetic children). The present study was approved by theReview Board of Kumamoto University School of Medicine. Weobtained the informed consent of the study subjects and their pa-tients.

Methods

Blood samples

Freshly drawn venous blood (10 ml) from subjects who had fastedovernight was collected into sterile tubes containing EDTA (®nalconcentration, 1 mg/ml), and plasma was isolated by low-speedcentrifugation (1000 g, 20 min at 4°C). Plasma samples were used inthe following experiments immediately after isolation.

Determination of LDL-size

LDL was isolated by single vertical-spin density-gradient ultra-centrifugation [35]. LDL-size was evaluated by electrophoresis innondenaturing polyacrylamide gradient gels on Pharmacia precastPAA 2/16 gels according to the procedure speci®ed by the manu-facturer. Standards used for size calibration purposes included la-tex beads (37 nm) (Dow Chemical Company) and Pharmacia high-molecular-weight standards (Pharmacia). The stained gels werescanned with a laser scanning densitometer (model CS-9000,Shimadzu) to provide a quantitative measurement of the size of thepeak and its distance from the origin. Particle diameter was cal-culated from a plot of the log of the known diameters of thestandards (latex beads 37 nm, thyroglobulin 17 nm, apoferritin12.2 nm) on the y-axis against their positions from the origin of thegel (Rf) on the x-axis.

Table 1 Clinical data on diabetic patients and controls (IBW ideal body weight)

Age (year)(range)

IBW(%)

Insulin dose(IU/kg) (range)

Fasting glucose(mg/dl)a

HbA1c(%)

MaleDiabetics (n=10) 13.5 � 4.8 (7±17) 106 � 3 0.8 � 0.2 (0.6±1.5) 135 � 25* 9.6 � 1.4*Controls (n=12) 14.5 � 3.1 (7±18) 109 � 4 0 88 � 8 5.0 � 0.5

FemaleDiabetics (n=13) 14.3 � 5.0 (7±18) 105 � 4 0.9 � 0.4 (0.6±1.7) 143 � 33* 8.7 � 1.3*Controls (n=15) 15.7 � 4.1 (8±19) 108 � 5 0 85 � 7 4.9 � 0.4

Values are expressed in mean � SD. *P<0.001: signi®cantly di�erent from controlsa To convert to mmol/l, multiply by 0.0556

277

Molar and fractional esteri®cation rates in plasma and very lowdensity lapoprotein (VLDL) and LDL-depleted plasma

The molar and fractional esteri®cation rates of cholesterol in plasma(MERplasma and FERplasma) and in VLDL and LDL-depleted plas-ma (MERHDL and FERHDL) were determined as described previ-ously [17, 18, 39].

Analysis of apolipoproteins, lipoproteins and lipids

ApoA-I, apoA-II, apoB, apoC-II, apoC-III and apoE concentra-tions in plasma were measured by radial immunodi�usion assay[39]. Concentrations of total cholesterol (TC), FC and TG in thesesamples were determined by enzymatic methods using commercialkits. HDL-C was measured by selective precipitation of LDL usingphosphotungstate-MgCl2 [8, 46]. The protein content of LDL wasdetermined by the method of Lowry et al. [29].

Statistical evaluation

Group di�erences were determined by Mann Whitney-U test.Group di�erences with P<0.05 were considered to be statisticallysigni®cant. Pearson correlation coe�cients were computed to as-sess the associations between parameters.

Results

Plasma lipids and apolipoprotein levels in childrenwith IDDM and controls

Plasma lipid and apolipoprotein levels are summarizedin Table 2. The plasma TG level was signi®cantly lowerin male and female diabetic children than in the re-spective controls (P<0.05 or 0.001). HDL-C andapoA-I levels in diabetic female children were signi®-cantly higher than those in controls (P<0.01). HDL-Clevel in male patients was signi®cantly higher than thatin controls (P<0.05). The apoA-II level was lower indiabetic children than in controls, but this di�erence wassigni®cant only among females (P<0.05). The ratio ofapoA-I to apoA-II in diabetic children was signi®cantlyhigher than that in controls (P<0.05 or 0.001). Other

lipid and apolipoprotein levels listed in Table 2 weresimilar between diabetic children and controls. HbA1C

level in diabetic children was not signi®cantly associatedwith the lipid and apolipoprotein parameters listed inTables 2 and 3. Age and duration of diabetes in diabeticchildren were also not associated with the parameterslisted in Tables 2 and 3.

Particle size of LDL in children with IDDMand controls

The distributions of LDL particle diameters in male andfemale diabetic children and controls are shown inFig. 1. Diabetic children were characterized by havingpredominantly large LDL particles: all had LDL-size>26.5 nm. The mean LDL particle diameters in maleand female diabetic children were signi®cantly greaterthan those in the respective controls (P<0.05) (Mean �SD in male diabetic children, male controls, female di-abetic children and female controls: 27.1 � 0.3, 26.5 �0.8, 27.3 � 0.4 and 26.8 � 0.6 nm). We previouslyreported [40] that FERHDL can predict LDL-size inaddition to the size distribution and physicochemicalcharacteristics of HDL in children and adolescents.Thus, we determined FERHDL in diabetic children. Asigni®cant correlation was not found between LDL-sizeand HbA1C level in diabetic children.

Fractional esteri®cation rate in plasmaand VLDL/LDL-depleted plasma in childrenwith IDDM and controls

The distribution of FERHDL in male and female diabeticchildren and controls are shown in Fig. 2. Diabeticchildren were characterized by having lower FERHDL:all had FERHDL <10%. The mean values for FERHDL

in male and female diabetic children were signi®cantlylower than those in the respective controls (P<0.01 or

Table 2 Plasma lipid and apo-lipoprotein levels (TC totalcholesterol, TG triglyceride,ApoA-I/ApoA-II the ratio ofapoA-I to apoA-II)

Male Female

Controls Diabetics Controls Diabetics

TCa 184 � 31 174 � 27 181 � 33 192 � 34TGb 100 � 35 56 � 24* 102 � 29 58 � 23***HDL-Ca 56 � 13 71 � 15* 59 � 12 72 � 12**ApoA-I 127 � 14 145 � 15 130 � 12 155 � 18**ApoA-II 34 � 5 32 � 5 31 � 2 29 � 3*ApoA-I/ApoA-II 2.4 � 0.5 2.8 � 0.3* 2.6 � 0.3 3.4 � 0.6***ApoB 88 � 15 74 � 20 85 � 17 73 � 19ApoC-II 3.8 � 1.5 3.1 � 1.4 3.3 � 1.1 2.7 � 1.1ApoC-III 7.8 � 2.6 8.6 � 2.6 8.3 � 2.2 8.2 � 2.7ApoE 5.5 � 2.6 4.3 � 1.0 5.7 � 1.1 5.2 � 0.8

Values are expresses in mean � SD (mg/dl) except for apoA-I/ApoA-IIa To convert to mmol/l, multiply by 0.0259bTo convert to mmol/l, multiply by 0.0113*P<0.05**P<0.01***P<0.001: signi®cantly di�erent from control

278

0.001) (Mean � SD in male diabetic children, malecontrols, female diabetic children and female controls:5.6 � 1.8, 8.1 � 2.1, 4.0 � 1.1 and 6.9 � 2.0%).FERplasma, MERplasma and MERHDL in diabetic chil-dren were all similar to those in the respective controls(Table 3). HDL-FC was signi®cantly higher in diabeticchildren than in controls, as expected by higher HDL-Clevels in diabetes. Signi®cant correlation was not foundbetween FERHDL and HbA1C level in diabetic children.

Discussion

Consistent with epidemiological data [3, 10, 12, 13],in vitro experiments have shown that small LDL(pattern B; particle diameter £ 25.5 nm) particles possessa lower binding a�nity for cellular LDL receptor andare more easily oxidized than large LDL (pattern A;particle diameter >25.5 nm), which suggests that small

Table 3 Fractional esteri®ca-tion rate in plasma and lowdensity lipoprotein/very lowdensity lipoprotein depletedplasma (FC free cholesterol)

FERplasma Total FCa MERplasma HDL-FC MERHDL

(%/h) (mg/dl) (lmol/h/l) (mg/dl) (lmol/h/l)

MaleDiabetics 2.3 � 0.4 41.8 � 6.5 20.2 � 10.5 12.5 � 3.6* 18.9 � 4.8Controls 2.3 � 0.4 41.3 � 7.1 24.1 � 4.4 8.6 � 4.0 16.5 � 2.9

FemaleDiabetics 2.2 � 0.5 46.4 � 6.3 19.8 � 11.6 14.1 � 3.3** 14.7 � 6.4Control 2.4 � 0.5 43.3 � 4.7 26.7 � 7.0 9.6 � 2.3 17.9 � 3.7

Values are expressed in mean � SD*P<0.05**P<0.01: signi®cantly di�erent from normala To convert mmol/l, multiply by 0.0259

Fig. 1 The distribution of LDLparticle size in diabetic childrenand controls

Fig. 2 The distribution ofFERHDL in diabetic children andcontrols

279

LDL particles are atherogenic [2, 14, 15, 45]. In thepresent study, LDL-size in diabetic children was greaterthan 26.5 nm. This suggests that LDL in diabetic chil-dren might be easily taken up by the liver, and may beresistant to oxidative modi®cation.

FERHDL re¯ects the reactivity of HDL to LCAT.Neither the LCAT concentration nor the substrate (freecholesterol and lecithin on HDL) is rate-limiting inFERHDL [17±19, 39]. The most important factor inFERHDL is the nature of the HDL particle substrate;i.e., the physicochemical characteristics of the particlesurface where LCAT interacts with its substrate [18, 19].Small HDL particles (HDL3b,c and small LpA-I parti-cles) provide the most active surface for LCAT reac-tions, while large HDL particles (HDL2b and large LpA-I particles) inhibit the reaction rate [5, 6, 25]. FERHDL ispositively correlated with plasma concentrations ofsmall HDL (atherogenic fraction) and negatively corre-lated with large HDL (anti-atherogenic fraction)[16, 17]. In diabetic children, FERHDL is signi®cantlylower than in nondiabetic controls. This suggests thatdiabetic children have large HDL particles than nondi-abetic controls. Based on the current concept of HDLand LDL, a lower FERHDL in diabetic children, inconcert with larger LDL particles, indicates that diabeticchildren may not be at high risk for CAD.

We previously reported that FERHDL shows a goodcorrelation with LDL-size in children [40]. The factorswhich regulate LDL-size are not completely understood.Several processes appear to be involved [1, 19, 23, 31];(1) hepatic production of apoB-containing lipoproteins[1]; (2) direct hepatic production of small dense LDL[1, 23, 31]; (3) intravascular lipolytic system [23]; and(4) the rate of the catabolism of LDL by an LDL recep-tor-dependent pathway [19]. Lipoprotein lipase (LPL),which is a key protein in the lipolytic system, is reduced insubjects with IDDM and NIDDM [41]. In general, themore severe the insulin de®ciency or resistance, thegreater the likelihood that decreases in LPL contribute todelayed TG metabolism (hypertriglyceridaemia). How-ever, after hyperglycaemia is treated by insulin, TG levelsdecrease substantially. LPL levels in postheparin plasmareturn to normal after 1 month of treatment [7, 41]. Ourdiabetic children had been treated with insulin for morethan 2 years and basal insulin levels (by exogenous insu-lin) were slightly higher than those in nondiabetic con-trols (data not shown). Thus, it seems likely that lipolysisin diabetic children is accelerated. Lower TG levels in ourdiabetic children might support this idea (Table 2).Therefore, accelerated lipolysis resulting in lower TGlevels is the most likely explanation for the change in bothLDL particle size and FERHDL in diabetic children. Asmentioned earlier, FERHDL is closely related to the par-ticle size distribution in HDL. Other workers havedemonstrated a relationship between LDL and HDLparticle size and the role of TGmetabolism in lipoproteinremodelling [33, 47]. Increased lipolysis may preventthe remodelling of HDL and increase large HDL par-ticles.

According to the current concept, the level of gly-caemic control is the major determinant of the lipopro-tein abnormalities in adult patients with IDDM [20]. Asmentioned earlier, IDDM patients with poor glycaemiccontrol usually have hypertriglyceridaemia caused bylowering LPL activity. However, limitations of intensiveinsulin therapy have been observed to achieve an ideallipoprotein metabolism (severe hypoglycaemic attacksoften occur to normalize lipoprotein metabolism) [20].Di�erent to adult patients with IDDM, plasma TGlevels were lower in our diabetic children than in con-trols although insulin dosage in our diabetic children isnot adequate for glycaemic control (HbA1C is stillhigh:8%±9%). This suggest that the e�ect of insulin onthe activity of LPL might be greater in children than inadults. Duration of diabetic condition may a�ect thee�ect of insulin on the LPL activity. Further studies areneeded to clarify this possibility.

In conclusion, we previously reported that plasmalevels of HDL and subspecies of HDL in diabetic chil-dren are quantitatively either normal or elevated [38].Now, we have show that qualitative changes in LDL andHDL in diabetic children are similar to those associatedwith a reduced risk for CAD. It appears that insulintherapy is more e�ective in diabetic children than indiabetic adults to correct lipoprotein abnormalities. Al-though CAD remains a major complication in patientswith IDDM even after the introduction of intensive in-sulin therapy [20], long-term intensive insulin therapy inchildren may help to prevent this complication.

Acknowledgements This study was supported in part by a ResearchGrant (7C-1) for Cardiovascular Disease from the Ministry ofHealth and Welfare, a Research Grant from Japan Study Groupfor Paediatric and Adolescent Diabetes and grants-in-aid(No. 09670817 and 09670773) from the Ministry of Education,Science and Culture of Japan.

References

1. Arad Y, Ramakrishnan R, Ginsberg HN (1990) Lovastatintherapy reduces low density lipoprotein apoB level in subjectswith combined hyperlipidemia by reducing the production ofapoB-containing lipoproteins: implications for the pathophysi-ology of apoB production. J Lipid Res 31:567±582

2. Arad Y, Rumsey SC, Galeano NF, Al-Haideri M, Walsh MT,Marcel YL, Milne RW, Kwiterovich PO, Deckelbaum RJ(1992) LDL has optimal size for maximal binding to the LDLreceptor. Circulation 86:I-551

3. Austin MA, Breslow JL, Hennekens CH, Buring JE, WillettWC, Krauss RM (1988) Low-density lipoprotein subclass pat-terns and risk of myocardial infarction. JAMA 260:1917±1921

4. Barakat HA, Carpenter JW, McLendon VD (1990) In¯uence ofobesity, impaired glucose tolerance, and NIDDM on LDLstructure and composition. Diabetes 39:1527±1533

5. Barter PJ, Hopkins GJ, Gorjatschko L, Jones ME (1984)Competitive inhibition of plasma cholesterol esteri®cation byhuman high density lipoprotein subfraction 2. Biochim BiophysActa 793:260±268

6. Barter PJ, Hopkings GJ, Gorjatschko L (1985) Lipoproteinsubstrates for plasma cholesterol esteri®cation. In¯uence ofparticle size and composition of the high density lipoproteinsubfraction 3. Atherosclerosis 58:97±107

280

7. Brunzell JD, Porte D Jr, Bierman EL (1975) Reversible ab-normalities in postheparin lipolytic activity during the latephase of release in diabetes mellitus (postheparin lipolytic ac-tivity in diabetes). Metabolism 24:1123±1137

8. Burstein M, Scholnick HR, Mor®n R (1970) Rapid method forthe isolation of lipoproteins from human serum by precipitationwith polyanions. J Lipid Res 11:583±595

9. Calvert GD, Graham JJ, Mannik T, Wise PH, Yeates EA(1978) E�ects of therapy on plasma high density lipoproteincholesterol concentration in diabetes mellitus. Lancet II:66±68

10. Campos H, Genest Jr JJ, Blijlevens E, McNamara JR, JennerJL, Ordovas JM, Wilson PWF, Schaefer EJ (1992) Low densitylipoprotein particle size and coronary artery disease. Ar-terioscler Thromb 12:187±195

11. Castelli WP, Garrison RJ, Wilson PWF, Abbott RD,Kalousdian S, Kannel WB (1986) Incidence of coronary heartdisease and lipoprotein cholesterol levels: The FraminghamHeart Study. JAMA 256:2835±2838

12. Coresh J, Kwiterovich Jr PO, Smith HH, Bachorik PS (1993)Association of plasma triglyceride concentration and LDLparticle diameter, density, and chemical composition with pre-mature coronary artery disease in men and women. J Lipid Res34:1687±1697

13. Crouse JR, Parks JS, Schey HM, Kahl FR (1985) Studies of lowdensity lipoprotein molecular weight in human beings withcoronary artery disease. J Lipid Res 26:566±574

14. De Graaf J, Hak-Lemmers HL, Hectors MP, Demacker PN,Hendriks JC, Stalenhoef AF (1991) Enhanced susceptibility toin vitro oxidation of the dense low density lipoprotein sub-fraction in healthy subjects. Arterioscler Thromb 11:298±306

15. Dejager S, Bruckert E, Chapman MJ (1993) Dense low densitylipoprotein subspecies with diminished oxidative resistance pre-dominate in combined hyperlipidemia. J Lipid Res 34:295±308

16. Dobiasova, Stribrna J, Sparks DL, Pritchard PH, Frohlich JJ(1991) Cholesterol esteri®cation rates in very low density lipo-protein- and low density lipoprotein-depleted plasma. Relationto high density lipoprotein subspecies, sex, hyperlipidemia, andcoronary artery disease. Arterioscler Thromb 11:64±70

17. Dobiasova M, Frohlich JJ (1994) Structural and functionalassessment of HDL heterogeneity. Clin Chem 40:1554±1558

18. Dobiasova M (1983) Lecithin cholesterol acyltransferase andthe regulation of endogenous cholesterol transport. Adv LipidRes 20:107±194

19. Eisenberg S, Gavish D, Oschry Y, Fainaru M, Deckelbaum RG(1984) Abnormalities in very low, low, and high density lipo-proteins and hypertriglyceridemia. Reverse toward normal withbeza®brate treatment. J Clin Invest 74:470±482

20. Garg A (1994) Management of dyslipidemia in IDDM patients.Diabetes Care 17:224±234

21. Genest Jr JJ, McNamara JR, Salem DN, Schaefer EJ (1991)Prevalence of risk factors in men with premature coronary ar-tery disease. Am J Cardiol 67:1185±1189

22. Gordon T, Castelli WP, Hjortland MC, Kannel WB, DawberTR (1977) High density lipoprotein as a protective factoragainst coronary heart disease. Am J Med 62:707±714

23. Grundy SM, Chait A, Brunzell JD (1989) Familial combinedhyperlipidemia workshop. Arteriosclerosis 7:203±207

24. Hamsten A, Walldius G, Szamosi A, Dahlen G, Faire U de(1986) Relationship of angiographically de®ned coronary arterydisease to serum lipoproteins and apolipoproteins in youngsurvivors of myocardial infarction. Circulation 73:1097±1110

25. Ikeda Y, Ohta T, Matsuda I (1994) Interaction between apoA-I-containing lipoproteins and lecithin: cholesterol acyltransfer-ase. Biochim Biophys Acta 1215:307±313

26. Kannel WB, Castelli WP, Goldon T, McNamara PM (1971)Serum cholesterol, lipoproteins and the risk of coronary heartdisease. Ann Intern Med 74:1±12

27. Kessler I (1971) Mortality experience of diabetic patients. Atwenty-six year follow-up study. Am J Med 51:715±724

28. Lopes-Virella MFL, Stone PG, Colwell JA (1977) Serum highdensity lipoproteins in diabetic patients. Diabetologia 13:285±291

29. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951)Protein measurement with Folin phenol reagent. J Biol Chem193:265±275

30. Marks HH, Krall LF (1971) Onset, course, prognosis andmortality in diabetes mellitus. In: Marble A, White P, BradleyRF, Krall LF (eds), Joslin's diabetes mellitus. Lea and Febiger,Philadelphia, pp 209±254

31. Marzetta AD, Foster DM, Brunzell JD (1990) Conversion ofplasma VLDL and IDL precursors into various LDL subpop-ulations using density gradient ultracentrifugation. J Lipid Res31:975±984

32. Miller GJ, Miller NE (1975) Plasma high density lipoproteinconcentration and development of ischemic heart disease.Lancet I:16±19

33. Murakami T, Michelangnoli S, Longhi R, Gianfranceschi G,Pazzucconi F, Calabresi L, Sirtori CR, Franceschini G (1995)Triglyceride are major determinants of cholesterol esteri®ca-tion/transfer and HDL remodeling in human plasma. Ar-terioscler Thromb Vasc Biol 15:1819±1828

34. National Diabetes Data Group (1979) Classi®cation and diag-nosis of diabetes mellitus and other categories of glucose in-tolerance. Diabetes 28:1039±1057

35. Nichols AV, Krauss RM, Lindgren FT, Forte TM (1986)Nondenaturing polyacrylamide gradient gel electrophoresis.Methods Enzymol 128:417±431

36. Nieminen MS, Mattila KJ, Aalto-Setala K, Kuusi T, KontulaK, Kauppinen-Makelin R, Ehnholm C, jauhiainen M, Valle M,Taskinen MR (1992) Lipoproteins and their genetic variation insubjects with and without angiographically veri®ed coronaryartery disease. Arterioscler Thromb 12:58±69

37. Norum KR, Gjone E, Glomset JA (1989) Familial lecithin:cholesterol acyltransferase de®ciency including ®sh eye disease.In: Scriver CR, Beaudet AL, Sly WS, Valle D (eds) The meta-bolic basis of inherited disease, 6th edn. McGraw-Hill, NewYork, pp 1181±1194

38. Ohta T, Nakamura R, Nishiyama S, Kodama M, Matsuda I(1990) Lipid and apolipoprotein compositions of two species ofapoA-I containing lipoproteins in young girls with insulin-de-pendent diabetes mellitus. Pediatr Res 28:42±45

39. Ohta T, Saku K, Takata K, Nakamura R, Ikeda Y, Matsuda I(1995) Di�erent e�ects of subclasses of HDL containing apoA-Ibut not apoA-II (LpA-I) on cholesterol esteri®cation in plasmaand net cholesterol e�ux from foam cells. Arterioscler ThrombVasc Biol 15:956±962

40. Ohta T, Kakiuti Y, Kurahara K, Saku K, Nagata N, Matsuda I(1997) Fractional esteri®cation rate of cholesterol in high den-sity lipoprotein is correlated with low density lipoprotein par-ticle size in children. J Lipid Res 38:139±146

41. O'Looney PA, Vahouny GV (1987) Diabetes and lipoproteinlipase activity. In: Borensztajn J (ed). Lipoprotein lipase.Evener, Chicago, pp 229±246

42. Rader DJ, Ikewaki K, Duverger N, Feuerstein I, Zech L,Conner W, Brewer Jr HB (1993) Very low high-density lipo-proteins without coronary atherosclerosis. Lancet 342:1455±1458

43. Reckless JPD, Betteridge DJ, Wu P, Payne B, Galton DJ (1978)High density and low density lipoprotein and prevalence ofvascular disease in diabetes mellitus. BMJ 1:883±886

44. Schaefer EJ, Zech LA, Schwartz DE, Brewer Jr HB (1980)Coronary heart disease prevalence and other clinical features infamilial high-density lipoprotein de®ciency (Tangier disease).Ann Intern Med 93:261±266

45. Tribble DL, Holl LG, Wood PD, Krauss RM (1992) Varia-tions in oxidative susceptibility among six low density sub-fractions of di�ering density and particle size. Arteriosclerosis93:189±199

46. Warnick GR, Nguyen T, Albers AA (1985) Comparison ofimproved precipitation methods for quanti®cation of highdensity lipoprotein cholesterol. Clin Chem 31:217±222

47. Williams PT, Karen KM, Krauss RM (1992) Correlations ofplasma lipoproteins with LDL subfractions by particle size inmen and women. J Lipid Res 33:765±774

281