885

Diabetes mellitus is a major risk factor for in-creased cardiovascular morbidity and mortalityrates.1 Diabetic cardiomyopathy was described as adisease entity in patients with insulin-dependent dia-betes mellitus in the absence of concomitant abnor-malities.2,3 Despite the known association betweendiabetes mellitus and increased cardiovascular mor-bidity and mortality rates, the prevalence of myocar-

dial systolic and diastolic functional abnormalities inasymptomatic patients with non–insulin-dependentdiabetes mellitus (NIDDM) is not well defined. Ad-ditionally, the correlation between such abnormali-ties and other diabetic complications such asnephropathy, retinopathy, and neuropathy is contro-versial.4-8 The aims of this work were (1) to assessleft ventricular (LV) diastolic and systolic function inpatients with NIDDM who have no symptoms orclinical evidence of cardiac disease and (2) to studythe relations among LV function and other specificdiabetic complications (retinopathy and autonomiccardiac neuropathy).

METHODS

The study population consisted of 66 patients withNIDDM who were selected from the patients attending the

Left Ventricular Systolic and DiastolicFunctional Abnormalities in Asymptomatic

Patients With Non–Insulin-DependentDiabetes Mellitus

Abdul Khaliq M. H. Annonu, MD, Alia Abdel Fattah, MD, M. Sherif Mokhtar, MD,Soliman Ghareeb, MD, and Abdou Elhendy, MD, PhD,

Cairo, Egypt, and Rochester, Minnesota

From the Departments of Cardiology and Critical Care Medicine,Cairo University Hospital, Cairo, Egypt; and the Division ofCardiovascular Diseases and Internal Medicine (A.E.), MayoClinic, Rochester, Minn.Reprint requests: Abdou Elhendy, MD, PhD, Mayo Clinic,Plummer Building A1, 200 1st St SW, Rochester, MN (E-mail:Elhendy.abdou@mayo.edu).Copyright © 2001 by the American Society of Echocardiography.0894-7317/2001/$35.00 + 0 27/1/112892doi:10.1067/mje.2001.112892

The aim of this study was to evaluate the relationsbetween left ventricular (LV) functional abnormali-ties, microangiopathy, and autonomic dysfunctionin patients with non–insulin-dependent diabetesmellitus (NIDDM). We studied 66 normotensivepatients with NIDDM of ≥4 years’ duration (age, 51 ±4.5 years; 35 men) and no clinical evidence of cardiacdisease. Twenty-one healthy subjects matched for ageand sex served as a control group. Echocardiographyand Doppler studies were performed to assess LVsystolic and diastolic function. Microangiopathy wasassessed by fundus examination. Autonomic functionwas assessed by standing blood pressure and heartrate response to Valsalva maneuver. Patients withNIDDM had a lower ejection fraction (58% ± 11%versus 66% ± 4%, P < .0001), E-F deceleration slope(382 ± 75 versus 427 ± 31 cm/s2, P < .05), and Evelocity (55 ± 11 vs 58 ± 6 cm/s, P = .02) of the mitraldiastolic flow, compared with control subjects, re-

spectively. Patients with ejection fraction <50% hada higher prevalence of retinopathy (65% versus29%, P < .005), abnormal blood pressure response tostanding (53% versus 8%, P < .0005), and proteinuria(65% versus 27%, P = .006). An inverse correlationwas found between the duration of diabetes and boththe ejection fraction (r = –0.53, P < .05) and E/A ratio(r = –0.4, P < .005). E/A ratio <1 was associated witha higher prevalence of retinopathy (49% versus 20%,P = .01) and abnormal blood pressure response tostanding (29% versus 4%, P < .005). Patients withNIDDM and no symptoms of cardiovascular diseasehave a reduced LV systolic and diastolic function ascompared with healthy subjects. LV systolic anddiastolic abnormalities are correlated with the dura-tion of diabetes and with other diabetic microangio-pathies such as diabetic retinopathy and neuropathy.(J Am Soc Echocardiogr 2001;14:885-91.)

Diabetic Center of Cairo University hospital betweenOctober 1994 and May 1997. Patients were enrolled in thestudy if they had NIDDM for at least 4 years, treated withoral hypoglycemic agents during their disease course.Themean age was 51 ± 6.8 years (range, 39 to 64).There were35 men and 31 women (group 1).The diagnosis of NIDDMwas established by current World Health Organization cri-teria.9 Twenty-one apparently healthy individuals (10 men,11 women) matched for age and sex served as a controlgroup (group 2).The mean age was 49 ± 4.5 years (range,40 to 57). All subjects gave an informed consent to partic-ipate in the study.The hospital ethics committee approvedthe study protocol. Exclusion criteria for the two groupswere alcoholism, known or discovered hypertension dur-ing examination, and clinical or electrocardiographic evi-dence of heart disease. Ischemic heart disease was clini-cally excluded by Rose Questionnaire10 and resting elec-trocardiogram according to the Minnesota Code.11 All sub-jects of the study underwent the following.

Laboratory Investigation

These included measurement of overnight fasting (12hours) and 2-hour blood sugar, fasting lipids profile (totalcholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglyceride levels),serum level of urea and creatinine, and urine analysis forproteinuria, which was measured by reagent strip.

Echocardiographic Studies

Two-dimensional and M-mode imaging was performedwith commercially available echocardiographic machines(Hewlett-Packard Sonos 1000), equipped with 2.5- and 3.5-MHz phased-array transducers. M-mode echocardiographywas used to measure cardiac dimensions and wall thickness.Fractional shortening was calculated as LV end-diastolic –end-systolic/end-diastolic dimensions × 100. LV ejectionfraction (EF) was calculated by means of the modifiedSimpson rule technique.

Doppler Studies

The pulsed Doppler sample volume was positioned at themitral leaflet tips. The following measurements weretaken: maximal early filling velocity E wave (cm/s) andmaximal late (atrial) filling velocity A wave (cm/s).The E/Aratio was derived. Deceleration of the early diastolic flow,determined as the slope of the descending limb of the Ewave (in cm/s2) was measured.

Diabetic patients were subjected to the following.

Evaluation of Cardiovascular Autonomic NerveFunction Tests

Two standard tests were performed to measure cardiovas-cular autonomic nerve function12: (1) Unmasking parasym-

Journal of the American Society of Echocardiography886 Annonu et al September 2001

pathetic dysfunction via heart rate response to Valsalvamaneuvers: The result was expressed as Valsalva ratio(longest R-R interval after maneuver (bradycardia)/shortestR-R interval during maneuver (tachycardia). A Valsalvaratio ≥1.2 was defined as normal, between 1.11 and 1.2 asborderline, and ≤1.1 as abnormal; (2) unmasking sympa-thetic dysfunction by blood pressure to standing. A fall insystolic blood pressure on erect position of ≤10 mm Hgwas defined as normal, 11 to 29 mm Hg as borderline, and≥30 as abnormal.

Fundus Examination

A direct ophthalmoscopy was performed by an experi-enced ophthalmologist unaware of echocardiographic orDoppler parameters, after dilating the patient’s pupils withtropicamide 1%. The retinopathy status was classifiedunder the following categories: (1) no diabetic retinopa-thy; (2) background diabetic retinopathy, defined as thepresence of one or more of the following:microaneurysmspunctuate or striate intraretinal hemorrhages, and hardexudates; (3) preproliferative diabetic retinopathy definedas soft exudates,venous beading, and intraretinal microvas-cular abnormalities; (4) proliferative diabetic retinopathycharacterized by neovascularization on or within one diskdiameter of the disk in extent.13

Statistical Analysis

Continuous data were presented as mean ± SD. Dichoto-mous variables were expressed as percentages. The un-paired t test (2-tailed) was used to assess differences be-tween continuous variables.The chi-square or Fisher exacttest was used to compare categoric variables betweengroups. Pearson correlation coefficient r values were usedto examine relations among variables. A value of P < .05was considered significant.

RESULTS

Clinical Characteristics

There were no significant differences between bothgroups regarding age and sex (by matching criteria).Diabetic patients had a larger body mass index (30.2± 4.2 versus 26.6 ± 2.6 kg/m2,P = .0004) and a higherresting heart rate (83.0 ± 8 versus 74 ± 5 bpm, P =.0001) compared with the control subjects.

Autonomic Function

In diabetic patients, systolic blood pressure responseto standing was abnormal in 13 (20%) patients, bor-derline in 17 (26%) patients, and normal in 36 (55%)patients.Valsalva ratio was normal in 46 (70%),abnor-

Journal of the American Society of EchocardiographyVolume 14 Number 9 Annonu et al 887

mal in 1 (1.5%), and borderline 19 (29%) diabeticpatients.

Fundus Examination

In diabetic patients, fundus examination revealedbackground diabetic retinopathy in 14 (21%) patients,preproliferative diabetic retinopathy in 8 (12%)patients, and proliferative diabetic retinopathy in 3(5%) patients.

Laboratory Data

No significant difference was detected between bothgroups with regard to the levels of serum choles-terol, low-density lipoprotein, and creatinine (Table1). Diabetic patients had lower levels of serum high-density lipoprotein and higher levels of triglyceridesand urea than did the control subjects.

Echocardiographic Findings

Systolic Parameters. Diabetic patients had a sig-nificantly higher mean left atrial diameter, LV end-systolic and end-diastolic dimensions, and diastolicseptal thickness compared with control subjects(Table 2). Mean ejection fraction and fractionalshortening were lower in diabetic patients.A largerproportion of diabetic patients had fractional short-ening <25%, ejection fraction <50%, and LV end-diastolic dimension >5.6 cm than did the controlsubjects. Segmental wall motion abnormalities weredetected in 1 diabetic patient and in none of thecontrol subjects.

Diastolic Parameters. Diabetic patients had alower mean E velocity and a lower E-F decelerationslope (Table 2). A larger proportion of diabeticpatients had a reversed E/A ratio (<1) and E-F slope<420 cm/s2 compared with control subjects. A trendwas found of a lower E/A ratio in diabetic patients.

Correlation Between Clinical Variables and LVFunction

Tables 3 and 4 demonstrate the correlation betweencardiac function and clinical variables (continuousand categoric variables, respectively). Ejection frac-tion correlated negatively with the duration of dia-betes, blood urea, fasting and 2-hour blood sugar,body mass index, and serum creatinine, whereas thecorrelation was less significant with age, low-densitylipoprotein, and total cholesterol.The E/A ratio wasinversely related with age, duration of diabetes, urea,and creatinine.

Diabetic Retinopathy

Diabetic retinopathy was present in 65% of patientswith EF <50% and in 29% of patients with EF ≥50%(P < .005).With regard to diastolic function, diabeticretinopathy was encountered in 49% of patients withE/A <1 and in 20% of patients with E/A ≥1 (P < .05,Table 4).

Autonomic Function

An abnormal or borderline R-R Valsalva ratio was notrelated to any of the systolic or diastolic abnormali-ties. Abnormal systolic blood pressure response tostanding was encountered in 53% of patients with EF

Table 1 Comparison of laboratory findings in diabeticpatients and control subjects

Laboratory Diabetic Control Pdata patients subjects value

Fasting blood sugar (mg/dL) 203 ± 51 95 ± 9 .00012-h blood sugar (mg/dL) 261 ± 56 113 ± 10 .0001Total cholesterol (mg/dL) 203 ± 48 198 ± 26 .6High-density lipoprotein 39 ± 10 47 ± 3 .0004

cholesterol (mg/dL)Low-density lipoprotein 122 ± 50 114 ± 25 .5

cholesterol (mg/dL)Triglycerides (mg/dL) 214.8 ± 60.4 185.1 ± 22.9 .03Urea 32 ± 6 28 ± 5 .02Creatinine 0.9 ± 0.2 0.9 ± 0.1 .5

Data are mean ± SD.

Table 2 Comparison of echocardiographic and Dopplerparameters in diabetic patients and control subjects

Echo/ Diabetic ControlDoppler patients subjects Pparameters (n = 66) (n = 21) value

LV end-diastolic 5.3 ± 0.6 4.9 ± 0.5 .005dimension (cm)

>5.6 cm 20 (30%) 0 (0.0%) .0001LV end-systolic 3.6 ± 0.8 3.0 ± 0.5 .002

dimension (cm)Fractional shortening (%) 33.0 ± 8.6 39 ± 7.5 .004

<25% 19 (29%) 1 (5%) .0001Ejection fraction (%) 58 ± 11 66 ± 4.2 .002

<50% 17 (26%) 0 (0.0%) .0001Left atrium (cm) 3.6 ± 0.4 3.1 ± 0.4 .0001Aorta (cm) 2.9 ± 0.5 2.8 ± 0.3 .3Interventricular septum (cm) 0.9 ± 0.1 0.8 ± 0.2 .04Posterior wall thickness (cm) 0.9 ± 0.1 0.9 ± 0.1 .5Diastolic parametersE (cm/s) 55 ± 10.6 58.4 ± 6.2 .02A (cm/s) 59 ± 9.5 57.5 ± 7.5 .5E/A 0.9 ± 0.2 1.0 ± 0.1 .07E/A <1 41 (62%) 8 (38%) .01E-F slope (cm/s2) 382 ± 75.1 427 ± 31 .03E-F slope <420 cm/s2 42 (64%) 9 (43%) .04

Data are mean ± SD, or number of patients (%).LV, Left ventricular.

<50% and in 8% of patients with EF ≥50% (P =.0003). With regard to diastolic function, abnormalsystolic blood pressure response was encountered in29% of patients with E/A <1 and in 4% of patientswith E/A ≥1 (P < .005).

Proteinuria was detected in 24 (36%) diabeticpatients and was associated with a higher preva-lence of reduced EF (65% versus 27%, P = .006);whereas no correlation was found with diastolicparameters (Table 4).

DISCUSSION

Cardiac involvement in diabetes mellitus covers awide spectrum, ranging from asymptomatic silentischemia to manifest infarction or evident heart fail-ure.The clinicopathologic entity called “diabetic car-diomyopathy”has been described in diabetic patientswho have no evidence of large-vessel disease, hyper-tension, or valvular heart disease.2 Considerabledebate exists regarding the exact nature and cause ofcardiac dysfunction attributable to diabetes mellitus.Recent research has indicated a generalized mem-brane defect, which may cause abnormalities of cal-cium metabolism in nerves, cardiac, and smoothmuscle as well as in endothelial cells.14 This may leadto the development of neuropathy, primary cardio-myopathy, and atherosclerosis in diabetics. Altera-tions in the myocardial cells and capillaries, causedby diabetes mellitus, may lead to myocardial cellinjury and interstitial fibrosis and, ultimately, to ven-tricular systolic and diastolic dysfunction.15 A con-

Journal of the American Society of Echocardiography888 Annonu et al September 2001

troversy exists in the literature regarding the relationbetween LV functional abnormalities in diabeticpatients and complications/or manifestations ofother organ involvement such as microangiopathy,autonomic neuropathy, and proteinuria.

Current Study

This study evaluated systolic and diastolic LV func-tion in 66 patients with NIDDM of more than 4-yearduration.Twenty-one healthy volunteers served as acontrol group. Subjects with hypertension and thosewith known or suspected coronary artery diseasewere excluded from the study because of the inde-pendent impact of these conditions on LV function.Our study showed that LV systolic and diastolic func-tional abnormalities are frequently encountered inasymptomatic patients with NIDDM.Compared withcontrol subjects, patients with NIDDM had a signifi-cantly lower EF, and 28% of these patients had areduced LV systolic function, as indicated by an EF<50% by echocardiography. Doppler indexes of LVfilling demonstrated that patients with NIDDM had alower diastolic function compared with control sub-jects, as indicated by a lower early filling (E) velocityand the high prevalence (62%) of E/A ratio of themitral diastolic inflow velocity <1.

There is a controversy among previous studiesregarding abnormalities of cardiac dimensions andglobal systolic function in diabetic patients.Attali etal3 reported an increased LV dimension in 9% of 49diabetic patients (26 with type I and 23 with type II).The EF was less frequently affected. Mustonen et al4

found no significant differences among patients withinsulin-dependent diabetes mellitus, NIDDM, andcontrol subjects regarding the resting EF. However,diabetic patients had more significant reduction ofEF at follow-up. Vanninen et al16 reported that frac-tional shortening was lower in 17 women with newlydiagnosed NIDDM, compared with healthy women.

Diastolic dysfunction was reported as a major fac-tor contributing to reduced exercise capacity inpatients with NIDDM.17 In our study, impaired LVdiastolic function, observed as reversed E/A ratio,was detected in 62% of diabetic patients comparedwith 38% in the control subjects (P < .01). E-F decel-eration slope was significantly lower in diabeticpatients than in the control subjects. Abnormalitiesof LV diastolic function have been reported inpatients with insulin-dependent diabetes mellitus18,19

as well as NIDDM.7 Di Bonito et al20 reported thatabnormalities of LV diastolic function occurred in 16patients evaluated early in the course of NIDDM,before the occurrence of microangiopathies.

Table 3 Relations between left ventricular systolic anddiastolic function and continuous variables among diabeticpatients

Ejection E/Afraction ratio

Clinical, laboratory parameters r P r P

Age –0.29 .02 0.38 .002Body mass index (kg/cm2) –0.43 .001 0.16 .2Duration of diabetes –0.53 .0 0.36 .003Fasting blood sugar –0.35 .004 0.18 .12-h blood sugar –0.39 .001 0.08 .5Total cholesterol –0.26 .04 0.21 .09High-density lipoprotein 0.22 .08 0.19 .1Low-density lipoprotein –0.25 .04 0.21 .09Triglyceride –0.16 .2 0.10 .4Urea –0.43 .0 0.44 .0Creatinine –0.37 .002 0.26 .03

Journal of the American Society of EchocardiographyVolume 14 Number 9 Annonu et al 889

Relation Between LV Function and Duration ofDiabetes

In our study, the duration of diabetes was signifi-cantly correlated with EF and E/A of LV diastolic fill-ing. The relation between LV dysfunction and theduration of diabetes among previous studies is con-troversial.Friedman et al6 found that diabetic patientswith a duration of disease more than 1 year hadlower EF than did control subjects. Conversely, Attaliet al3 found that increased LV dimensions and de-creased EF were not related to duration of diabetesor to the presence or extent of complications.

Diabetic Complications

In this study, LV systolic and diastolic functionalabnormalities were correlated with other diabeticcomplications in the same population, whereas mostof the previous studies focused on either the systolicor the diastolic function and reported correlationwith one but not all of the major diabetic complica-tions.

Diabetic Retinopathy

Diabetic retinopathy is a predictor of microangiopa-thy in organs other than the eyes. Many studiesshowed a significant correlation between retinopa-thy and some of the chronic diabetic (microangio-pathic) complications such as nephropathy andneuropathy.21 In our study, diabetic retinopathy waspresent in 38% of patients with NIDDM.There wasa significant correlation between diabetic retinopa-thy with diminished EF (P < .005), reversed E/A ratio(P < .05), and diminished EF deceleration slope (P <.005). This may be explained by the associatedmicrovascular disease involving the coronary circu-lation with subsequent impairment of systolic anddiastolic LV performance. Akasaka et al22 reportedthat coronary flow reserve was significantly restrict-

ed in patients with diabetes mellitus, and its reduc-tion was more marked in those with diabetic retino-pathy, especially in advanced retinopathy. Hiramatsuet al23 reported that diabetic patients with retinopa-thy had significantly greater isovolumic relaxationtime and A/E values than those without retinopathy.In contrast with these findings, Zarich et al8 foundthat there is no relation between the LV diastolic fill-ing abnormalities with retinopathy, nephropathy, orperipheral neuropathy.

Autonomic Function

Autonomic cardiac neuropathy is associated with anincreased cardiovascular risk in diabetic patients.24

Di Carli et al25 reported that diabetic autonomic neu-ropathy was associated with an impaired vasodilatorresponse of coronary resistance vessels to increasedsympathetic stimulation, which was related to thedegree of sympathetic nerve dysfunction. Cardio-vascular autonomic nerve dysfunction was evaluatedin our study by the use of response to Valsalvamaneuver and blood pressure response to standing.Abnormal systolic blood pressure response to stand-ing was significantly correlated with diminished EFand reduced E/A ratio. In contrast to sympatheticdysfunction, the parasympathetic dysfunction hadno significant correlation with diminished LV systolicor diastolic performance in our study. Mustonen etal26 reported that the Valsalva ratio and heart ratevariation during deep breathing were lower in dia-betic patients than in normal subjects. An inversecorrelation was found between autonomic nervefunction score and LV diastolic filling, whereas nocorrelation was found with systolic function. Kahnet al7 reported that in 28 patients with insulin-dependent diabetes mellitus without evidence ofischemic heart disease, 21% had abnormal diastolicfilling and differed from diabetic patients with nor-

Table 4 Relations between left ventricular systolic and diastolic function and categoric variables among diabetic patients

Ejection fraction E/A

<50% ≥50% <1 ≥117 Patients 49 Patients P value 41 Patients 25 Patients P value

Blood pressure response to standingAbnormal 9 (53) 4 (8) .0003 12 (29) 1 (4) .003Borderline 5 (29) 12 (26) .5 9 (22) 8 (32) .5

Valsalva ratioAbnormal 1 (6) 0 (0.0) .3 1 (2) 0 (0.0) .6Borderline 6 (35) 13 (27) .4 8 (20) 11 (40) .09

Diabetic retinopathy 11 (65) 14 (29) .003 20 (49) 5 (20) .01Proteinuria 11 (65) 13 (27) .006 15 (37) 9 (36) .6

Data are presented as number of patients (%).

mal filling in their greater severity of cardiac auto-nomic neuropathy.

Proteinuria

Proteinuria is a strong predictor of vascular morbidityand mortality in patients with NIDDM.27 In ourstudy, gross proteinuria was detected in 36% ofpatients with NIDDM, despite that the mean ureaand creatinine levels were normal. Ejection fractionwas more frequently reduced among patients withproteinuria, whereas no correlation was foundbetween proteinuria and diastolic function.

Limitations of the Study

Pulmonary venous flow was not recorded, and Dop-pler studies were not repeated during the Valsalvamaneuver.Therefore pseudonormalization of LV fill-ing could not be studied. Coronary angiography wasnot performed in diabetic patients to exclude thepresence of coronary artery disease as the underly-ing cause of LV dysfunction. Finally, follow-up wasnot performed to evaluate the course and the long-term impact of these abnormalities.

Conclusions

Patients with NIDDM and no symptoms or signs ofcardiovascular disease have a reduced LV systolicand diastolic function as compared with healthy sub-jects. Left ventricular systolic abnormalities in NIDDMare correlated with the duration of diabetes and withother diabetic microangiopathies, as manifested bydiabetic retinopathy and autonomic nervous systemdysfunction. Diabetic retinopathy, abnormal systolicblood pressure response to standing, and the dura-tion of diabetes are predictive of LV diastolic dys-function.

REFERENCES

1. Stamler J, Vaccaro O, Neaton JD, Wentworth D. Diabetes,other risk factors, and 12-yr cardiovascular mortality for menscreened in the Multiple Risk Factor Intervention Trial.Diabetes Care 1993;16:434-4.

2. Galderisi M, Anderson KM, Wilson PW, Levy D. Echocardio-graphic evidence for the existence of a distinct diabetic car-diomyopathy (the Framingham Heart Study). Am J Cardiol1991;68:85-9.

3. Attali JR, Sachs RN, Valensi P, Palsky D, Tellier P, VulpillatM, et al. Asymptomatic diabetic cardiomyopathy: a noninva-sive study. Diabetes Res Clin Pract 1988;4:183-90.

4. Mustonen JN, Uusitupa MI, Laakso M, Vanninen E,Lansimies E, Kuikka JT, et al. Left ventricular systolic func-tion in middle-aged patients with diabetes mellitus. Am JCardiol 1994;73:1202-8.

5. Ferraro S, Perrone-Filardi P, Maddalena G, Desiderio A,

Journal of the American Society of Echocardiography890 Annonu et al September 2001

Gravina E, Turco S, et al. Comparison of left ventricular func-tion in insulin- and non-insulin-dependent diabetes mellitus.Am J Cardiol 1993;71:409-14.

6. Friedman NE, Levitsky LL, Edidin DV, Vitullo DA, LacinaSJ, Chiemmongkoltip P. Echocardiographic evidence forimpaired myocardial performance in children with type I dia-betes mellitus. Am J Med 1982;73:846-50.

7. Kahn JK, Zola B, Juni JE, Vinik AI. Radionuclide assessmentof left ventricular diastolic filling in diabetes mellitus with andwithout cardiac autonomic neuropathy. J Am Coll Cardiol1986;7:1303-9.

8. Zarich SW, Arbuckle BE, Cohen LR, Roberts M, Nesto RW.Diastolic abnormalities in young asymptomatic diabeticpatients assessed by pulsed Doppler echocardiography. J AmColl Cardiol 1988;12:114-20.

9. World Health Organization. Diabetes mellitus: report of aWHO Study Group. Geneva: World Health Organization;1985 (Tech Rep Ser No. 727).

10. Cook DG, Shaper AG, MacFarlane PW. Using the WHO(Rose) angina questionnaire in cardiovascular epidemiology.Int J Epidemiol 1989;18:607-13.

11. Prineas RJ, Crow RS, Blackburn H. The Minnesota CodeManual of Electrocardiographic Findings. Boston: JohnWright PSB; 1982:203-20.

12. Ewing DJ, Clarke BF. Autonomic neuropathy: its diagnosisand prognosis. Clin Endocrinol Metab 1986;15:855-88.

13. Chen MS, Kao CS, Chang CJ, Wu TJ, Fu CC, Chen CJ, etal. Prevalence and risk factors of diabetic retinopathy amongnoninsulin-dependent diabetic subjects. Am J Ophthalmol1992;114:723-30.

14. Pierce GN, Russell JC. Regulation of intracellular Ca2+ inthe heart during diabetes. Cardiovasc Res 1997;34:41-7.

15. Kawaguchi M, Techigawara M, Ishihata T, Asakura T, SaitoF, Machara K, et al. A comparison of ultrastructural changeson endomyocardial biopsy specimens obtained from patientswith diabetes mellitus with and without hypertension. HeartVessels 1997;12:267-74.

16. Vanninen E, Mustonen J, Vainio P, Lansimies E, Uusitupa M.Left ventricular function and dimensions in newly diagnosednon-insulin-dependent diabetes mellitus. Am J Cardiol1992;70:371-8.

17. Poirier P, Garneau C, Bogaty P, Nadeau A, Marois L, BrochuC, et al. Impact of left ventricular diastolic dysfunction onmaximal treadmill performance in normotensive subjects withwell-controlled type 2 diabetes mellitus. Am J Cardiol 2000;85:473-7.

18. Albanna II, Eichelberger SM, Khoury PR, Witt SA, Standi-ford DA, Dolan LM, et al. Diastolic dysfunction in youngpatients with insulin-dependent diabetes mellitus as deter-mined by automated border detection. J Am Soc Echo-cardiogr 1998;11:349-55.

19. Riggs W, Transue D. Doppler echocardiographic evaluationof left ventricular diastolic function in adolescents with dia-betes mellitus. Am J Cardiol 1990;65:899-902.

20. Di Bonito P, Cuomo S, Moio N, Sibilio G, Sabatini D,Quattrin S, et al. Diastolic dysfunction in patients with non-insulin-dependent diabetes mellitus of short duration. DiabetMed 1996;13:321-4.

21. Zander E, Heinke P, Herfurth S, Reindel J, Ostermann FE,Kerner W. Relations between diabetic retinopathy and car-diovascular neuropathy: a cross-sectional study in IDDM andNIDDM patients. Exp Clin Endocrinol Diabetes 1997;105:319-26.

Journal of the American Society of EchocardiographyVolume 14 Number 9 Annonu et al 891

22. Akasaka T, Yoshida K, Hozumi T, Takagi T, Kaji S, Kawa-moto T, et al. Retinopathy identifies marked restriction ofcoronary flow reserve in patients with diabetes mellitus. J AmColl Cardiol 1997;30:935-41.

23. Hiramatsu K, Ohara N, Shigematsu S, Aizawa T, Ishihara F,Niwa A, et al. Left ventricular filling abnormalities in non-insulin-dependent diabetes mellitus and improvement by ashort-term glycemic control. Am J Cardiol 1992;70:1185-9.

24. Maser RE, Pfeifer MA, Dorman JS, Kuller LH, Becker DJ,Orchard TJ. Diabetic autonomic neuropathy and cardiovas-cular risk: Pittsburgh Epidemiology of Diabetes Complica-tions Study III. Arch Intern Med 1990;150:1218-22.

25. Di Carli MF, Bianco-Batlles D, Landa ME, Kazmers A,Groehn H, Muzik O, et al. Effects of autonomic neuropathyon coronary blood flow in patients with diabetes mellitus.Circulation 1999;100:813-9.

26. Mustonen J, Mantysaari M, Kuikka J, Vanninen E, Vainio P,Lansimies E, et al. Decreased myocardial 123I-meta-iodobenzylguanidine uptake is associated with disturbed leftventricular diastolic filling in diabetes. Am Heart J 1992;123:804-5.

27. Agardh CD, Agardh E, Torffvit O. The prognostic value ofalbuminuria for the development of cardiovascular diseaseand retinopathy: a 5-year follow-up of 451 patients with type2 diabetes mellitus. Diabetes Res Clin Pract 1996;32:35-44.