Coronary Microangiopathy in Type 2 Diabetic Patients: Relation to Glycemic Control, Sex, and Microvascular Angina Rather Than to Coronary Artery Disease

Ikuo Yokoyama, Katsunori Yonekura, Tohru Ohtake, Weidong Yang, Wee Soo Shin, Nobuhiro Yarnada, Kuni Ohtomo, and Ryozo Nagai

Departments ofCardiovascular Medicine, Radiology, and Metabolic Diseases, University ofTokyo, Tokyo,Japan

thy in diabetics (7—15).However, the specific role of these Coronary microangiopathyis a major complicationin diabetics. factorsincoronarymicroangiopathyisunclear. However,thepresenceofindependentfactorsinassociationwith This coronary microangiopathy is often expressed as coronarymicroangiopathyin patientswith non—insulin-depenmicrovascular angina that is characterized by an angina with dent diabetesmellitus(NIDDM) or the differencein coronary normal coronary angiography. The association of hyperinsu microangiopathybetween diabeticswithcoronaryartery disease linemia in nondiabetics with microvascular angina without (CAD)andthosewithmicrovascularanginaisunclear.Methods: Nineteenpatientswith NIDDM and microvascularangina, 18 coronary artery disease (CAD) has been reported (16). patients with NIDDM and CAD, and 17 age-matched control However, the role of insulin resistance in such patients has not subjectswere studied. Myocardialsegments that were perfused been clarified.Althoughcoronarymicroangiopathyin diabetics by angiographicallynormalcoronaryarterieswerestudied.The has been studied, including that in patients with microvascular baseline myocardial blood flow (MBF) and the MBF during angina or with CAD (4,5), it has not been determined whether dipyridamole administration were measured using PET and differences exist in the nature of the myocardial perfusion 13N-ammonia,after which the myocardial flow reserve (MFR) abnormality between these 2 groups of patients. was calculatedto assesscoronarymicroangiopathy.Results: The baselineMBFwascomparableamongNIDDMpatientswith Recently, the relationship between glycemic control and microvascular angina, NIDDM patients with CAD, and control reduced myocardial flow reserve (MFR) in patients with non subjects.However,the MBFduringdipyndamoleadministration insulin-dependentdiabetesmeffitus(NIDDM)withoutevidence was significantly lower in NIDDM patients with microvascular of myocardial ischemia has been reported (17,18). Because angina (126 ±42.7 mLimin/100 g) than that in either NIDDM coronaryangiographycouldnotbe performedin thoseasymptom patients with CAD (210 ±70.1 mL/min/100 g; P < 0.01) or atic patients,it is not known whether the reduced MFR in such controlsubjects(293 ±159 mL/min/100 g;P < 0.01),aswasthe patients is associated with coronary microangiopathy or macroan MFR (NIDDM withmicrovascularangina,1.90 ±0.73; NIDDM with CAD, 2.59 ±0.81 [P < 0.01]; controlsubjects,3.69 ±1.09 giopathy (17,18). Which factors are responsible for the coronary [P < 0.01]). Multivanate stepwise regression analysis showed microangiopathy in NIDDM also remains uncertain. that,amongthefactors considered,glycemiccontrolwas indepen This study was undertaken to determine the factors that dently related to the MFR (r = 0.838; P < 0.05). ConclusIon: are the principalcontributorsof the reducedMFR in Glycemiccontrolappearstobeessentialforcoronarymicroangi angiographically normal coronary arteries in patients with opathyinNIDDM. NIDDM. Furthermore, we compared the MF'R of patients Key Words: hyperglycemia;non—insulin-dependentdiabetesmel with NIDDM having microvascular angina with the MFR of litus;flow reserve;atherosclerosis;coronarymicroangiopathy those having CAD, who were perfused by normal coronary J NucIMed2000;41:978-985 arteries.

MATERIALS AND METHODS oronary microangiopathy is an important pathophysi StudyPopulation ologic feature of diabetes (1—6). Insulin resistance, glycemic Thirty-seven patients with NIDDM (28 men, 9 women; mean control, and lipid disorders are related to vascular angiopa age, 62.0 ±8.08 y; age range, 42—68y) with microvascular angina (n = 19) or CAD (n = 18) were studied. Seventeen normolipid emic, nonnoglycemic, asymptomatic age-matched volunteers (13 ReceivedJun. 15, 1999; revision accepted Oct. 4, 1999. men, 4 women; mean age, 55.9 ± 10.4 y; age range, 38—68 y) Forcorrespondenceor reprintscontact:IkuoYokoyama,MD, Departmentof Cardiovascular Medicine, University ofTokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, without a history of heart disease or chronic disease were selected Japan113-8655. as control subjects. Diagnosis of microvascular angina was made

978 Ti-ii@Joui@i OF NUCLEARMEDICINE•Vol. 41 •No. 6 •June 2000 on thebasisofthe following criteria:positivefindingsbyelectroca.r specialists (0% stenosis). 1\velve NIDDM patients were treated diographicexercisestress testingand anatomicallynormalcoro with oral hypoglycemic agents, 21 patientswere treatedwith diet nary arteries. Details ofcoronary angiographic findings are given in therapyalone,and4 patientswho hadapparentlydeveloped insulin Table 1, and general characteristics of the study subjects are resistance were treated with insulin. The latter 4 patients had summarized in Table 2. There were no significant differences become resistant to oral hypoglycemic agents after long-term among the 3 groups in terms of sex, body weight, height, body therapywith these agents.Among the 37 patientswithNIDDM, 19 mass index (BMI), blood pressure at rest and during dipyridamole with well-controlled essential hypertensionwere included, 10 of administration, heart rate at rest and during dipyridamole loading, whom had diffuse left ventricular hypertrophy ([LVH]; wall and levels of plasma total cholesterol, low-density lipoprotein thickness, >12 mm). The criteria of essential hypertension were (LDL)cholesterol,high-densitylipoprotein(HDL)cholesterol,and systolic blood pressure(SBP) > 160 mm Hg or diastolic blood triglyceride.Plasmafastinginsulinconcentrationwascomparable pressure (DBP) > 95 mm Hg (or both) with no specific cause. Of between subjects with NIDDM and microvascular angina without the 10 patients with LVH, 8 had microvascularangina. These CAD and those with NIDDM and CAD There was no significant patients were not included in other recent studies (1 7—21).The differencein age betweentheage-matchedcontrolsubjects(control criterionof well-controlled diabetes was a HbA1@level averaged group) and the total group of patients with NIDDM or those with over the previous 5 y of <8.0% and that of poorly controlled NIDDM and microvascularangina. However, the mean age of diabetes was a HbA1@level averaged over the previous 5 y of patients with NIDDM and CAD was significantly higher than that 8.0%. of both control subjects and patients with NIDDM and microvascu Control Group. The resting electrocardiogram was normal in all lar angina. Plasma concentrationsof fasting glucose (FBS) and control subjects.All potentialcontrol subjectsunderwenta symp hemoglobin A1@(HbA1@) levels averaged for the past 5 y from data tom-limited treadmill test, and those with typical chest pain or gathered at regular clinic visits for diabetes were significantly abnormalelectrocardiogramindicatingmyocardialischemia were higher in those with NIDDM than in control subjects. Data for excluded. In this study,cardiacnormalityin control subjects was control subjects were reviewed from those collected at visits to not confirmedby coronaryangiography,but these subjects were local health service centers. HbA1@was assayed by the latex selected among those with a low probability of cardiac disease and aggregation method (normal range, 4.3%—5.8%).FBS was mea can be consideredto be appropriateas healthycontrol subjectsas sured after an overnight fast of >12 h. Plasma insulin concentra reported by Rozanski et al (22). Because of the difficulty in tions were measured by radioimmunoassay (normal fasting range, recruitinghealthy control subjects for PET studies, PET data for 3—18U/mL). Before the study, all participants were informed of the all but 2 of these control subjects were those used for other natureof the investigation,afterwhich they agreedto participatein studies (17—21). the study protocol, which was approved by the local ethics committee. EstImation of Insulin Resistance Patients with NIDDM. We made the diagnosis of NTDDM Quantitativeestimation of whole-body insulin resistance was according to the following criteria: fasting glucose concentration> made by obtaining the glucose disposal rate (GDR) during 7.3 mmol/L (140 mg/dL) and HbA1@level > 6.5% before the hyperinsulinemic euglycemic clamping (@.tmol/min/kg) using a initiation of therapy. All patients with NIDDM underwent coronary previously reported method (18). In all patients, the glucose angiography and were proven to have I or 2 normal coronary concentrationbecame constant within 2 h after the initiation of arteries within the 3 major branches as diagnosed by 3 independent insulinclamping. Estimationof MajorComplicationsof Diabetes TABLE 1 The presence and degree of retinopathy,neuropathy,and ne CoronaryAngiographicFindingsinStudyPatients phropathy,which aremajorcomplicationsof diabetes,were graded as follows after a review of clinical records: 0, none; 1, mild; 2, Coronaryangiographicfindings No. moderate;and3, severe. Zero-vesseldisease PET Microvascularangina 19 6 The regional baselinemyocardial blood flow (MBF) and the AfterPTCAto LAD AfterPTCAto LCX 1 MBF duringdipyridamoleloadingwere measuredusingPET and AfterPTCAto RCA 1 ‘3N-ammonia.Twenty-fourhoursbeforethePET study,all medica AfterCABGtoLADandLCX 1 tions were withheld and caffeine intake was stopped because One-vesseldisease caffeine intake can alter the MBF (23). Cigarette smoking was LAD 2 stop@ onthedayofthePETstudybecausecigarettesmokingcan LCX 1 reducetheMFR (24).Myocardialflowimageswereobtainedusing RCA 1 a Headtome IV scanner(Shimadzu Corp., Kyoto, Japan). This Two-vesseldisease 0 scannerhas7 imagingplanes;in-planeresolutionis 4.5 mm at full LADandLCX 3 width at half maximum (FWHM) and the z-axial resolutionis 9.5 LADand RCA LCXandRCA 3 mm at FWHM. Effective in-planeresolutionis 7 mm after usinga No. of patientswitholdmyocardialinfarction 3 smoothingfilter.The sensitivitiesof the HeadtomeIV scannersare 14and24kcts/sfordirectandcrossplanes,respectively. After acquiringtransmissiondatato correctfor photonattenua PTCA= percutaneoustransluminalcoronaryangioplasty;LAD= tion before obtainingimages, 740—1110MBq ‘3N-ammoniawere leftanteriordescendingartery;LCX = leftcircumflexartery;RCA = injected, and dynamic PET and static PET were performedfor 2 rightcoronaryartery CABG = coronaryarterybypassgrafting. and8 mm, respectively.Afterwaiting45 mm to allow fordecay of

CORONARY MIcRo@GloPAmY IN NIDDM •Yokoyama et al. 979 TABLE 2 GeneralCharacteristicsof StudySubjects

Control patientsMVA Characteristic subjects WithNIDDM WithCAD All

n(M1@ 17(13/4) 19 (16/3) 18 (12/6) 37 (28/9) Age (y) 55.9±10.4 58.9 ±8.57 65.3 ±6.33* 62.0 ±8.08 BW (kg) 60.7 ±7.81 65.6 ±11.6 62.1 ±8.75 65.2 ±9.22 Height(cm) 164 ±11.3 166 ±7.60 160 ±8.63 163±8.60 BMI 23.4 ±4.10 24.5 ±2.7 24.3 ±2.95 24.5 ±2.76 At rest SBP (mm Hg) 129 ±12.3 132 ±19.8 144 ±19.1 139±20.2 DBP (mm Hg) 76.8 ±7.34 77.7 ±8.60 77.7 ±13.2 79.0 ±10.5 HR (bpm) 67.1±13.9 67.5 ±8.31 60.3 ±12.7 65.2 ±10.8 APP 8892±1692 9017 ±2075 8523 ±1732 9067 ±1856 AfterDP SBP(mmHg) 123 ±11.7 129±19.5 135±24.1 134 ±20.6 DBP (mm Hg) 76.4 ±7.40 71.1 ±9.42 71.0 ±16.8 71.2 ±13.2 APP 9564 ±1012 10,423 ±2366 9427±1420 9673±1578 FBS (mmotfL) 4.82 ±0.49 10.4 ±2.50t 9.00 ±2.24t 9.72 ±2.37t HbA1@(%) 5.51±0.23 8.40 ±1.lOt 7.90 ±1.80t 8.3 ±1.4t TC (mmoVL) 5.01±0.56 4.62 ±0.727 4.89 ±0.740 4.77 ±0.694 HDL(mmoVL) 1.46 ±0.66 1.06 ±0.240 1.14±0.266 1.10 ±0.27 TG(mmoLIL) 1.28 ±0.45 1.37 ±0.906 1.47 ±0.959 1.51 ±0.97 LDL(mmovL) 2.99 ±0.50 2.84 ±1.03 3.07±0.728 2.89 ±0.89 Fl (mlU/L) 8.07 ±5.03 6.50 ±6.98 7.29 ±6.32

*P< 0.01 versusMVA. tP < 0.01 versuscontrols. MVA= microvascularangina;n = numberofstudypatients;BW= bodyweight;BMI= bodymassindex;SBP= systolicbloodpressure; DBP = diastolicbloodpressure;HR = heartrate;APP = ratepressureproducts;DP = dipyridamole;FBS = fastingplasmaglucose concentration;HbA1@= hemoglobinA1@;TC = total cholesterol;HDL = high-densitylipoproteincholesterol;TG = triglycerides;LDL = low-densitylipoproteincholesterol;Fl = fastingplasmainsulinconcentration.

the radioactivityof ‘3N-ammonia,dipyridamole(0.56 mg/kg) was raphy. The recovery coefficients were taken into consideration administrated intravenously. Five minutes after dipyridamole admin when measuringthe MBF. istration, 740—1110MBq ‘3N-ammoniawere injected and, at To determinethe MBF, regions of interest were placed at the exactly the same time, a second dynamic scan was obtainedfor 2 septum,anteriorwall, lateralwall, and inferoposteriorwall on the mm and a static scan was obtained for 8 mm. The dynamic scan transaxial images as reported (17,19,20). In this investigation, only was obtained every 15 s (8 times) during the 2-mm period, and segmentsthatwere perfusedby angiographicallynormalcoronary dynamicdatawereobtainedfor7 slices. Only 1-channelelectrocar arteries were studied. Segments that were perfused by coronary diographic monitoring in limb leads was made during PET arterybypass graftswere excluded because reducedMRF in such scanning. segmentshas been reported(28). Static 13N-ammoniaimages were Determination of MBF and MFR also obtained from PET and analyzed visually by 3 independent The regional MBF was calculatedaccordingto the 2-compart specialistswhohadnootherinformationon thepatients.TheMFR mentmodel (25,26). This modelhasbeen validated(25,26) andhas was determined as follows: MFR = regional MBF during DP been used frequentlyin studiesof MBFand MFR(17—21,23,24).administration/regionalbaselineMBF. Metabolitesof ‘3N-ammoniacan be negligible duringthe first90 s afterinfusionof 13N-ammonia(27).Thetime—activitycurveof the StatistIcs left ventricular cavity was used as an input function. Tracer TheMBFatrest,MBFduringdipyridamoleloading,MFR,body spillover was corrected by least-squares nonlinear regression analysisto calculatethe MBF withthe assumptionthatmyocardial weight, SBP, DBP, height, BMI, and lipid parameters were radioactivity and left ventricular radioactivity were influenced by comparedin the 3 groupsusingANOVA;individualdatawerethen each other.Details aregiven in recentarticles(17,19,20). analyzed by the 2-tailed unpaired Student t test. Values are All datawere correctedfor dead-timeeffects to reduceerrorto expressed as the mean ±SD. P < 0.05 was considered significant. <1%. Toavoidtheinfluenceofthe partial-volumeeffect associated Multivariatestepwiseregressionanalysiswasundertakento exam with the object's size, recovery coefficients obtained from experi inc which factors were independently related to MFR: age, baseline mentalphantomstudiesin our laboratorywereused.Therecovery MBF, average FBS, average HbA1@,FBS at the time of PET, total coefficient was 0.8 when myocardial wall thickness was 10 mm. cholesterol, triglyceride,HDL cholesterol, LDL cholesterol, dura For correction of the partial-volume effect, specialists in our tion of diabetic state, plasma insulin concentrationduring PET, hospital measured wall thickness with 2-dimensional echocardiog GDR,andtheexistenceof microvascularangina.

980 THE JOURNALOF NUCLEARMEDICINE•Vol. 41 •No. 6 •June 2000 RESULTS microvascular angina (n = 16; 1.79 ±0.62) was signifi HemodynamicandElectrocardiographicResponsesto cantly lower than that in men with NIDDM and CAD (n = DipyridamoleAdministration 12; 2.43 ±0.34; P < 0.05), although the patients with No significant differences in SBP at rest and during NIDDM and CAD were significantly older. However, no dipyridamole administration and rate pressure product (RPP) such difference was observed in female NIDDM patients. were found between the groups (Table 2). During dipynda The MFR (1.52 ±0.436) in 10 NIDDM patients with mole administration, typical chest pain or chest oppression LVH was significantly lower than that in NIDDM patients accompanied by electrocardiographic changes was observed without LVH (2.17 ±0.69; P < 0.05), whereas the baseline in 24 study patients. Because of difficulty in recording the MBF in NIDDM patients with LVH (105 ±27.6) was electrocardiogramin the precordial leads during PET, a significantly higher than that in NIDDM patients without detailed description of electrocardiographic response to LVH (81.4 ±14.6; P < 0.01). However, the MBFs during dipyridamole was not possible. dipyridamole administration were comparable between those with and without LVH (158 ±50.7 versus 182 ±57.6). BaselineMBF,MBFDuringDipyridamole Administration,andMFR Furthermore, the percentage of patients with LVH was higher among those with microvascular angina (8/19; 47.4%) The baseline MBF was comparable among patients with than those with the CAD group (2/20; 10%). NIDDM having microvascular angina (78.7 ±18.8 mUmin/ 100 g), those having CAD (81.8 ±16.2 mL/min/l00 g), and MFR and Hypertension control subjects (78.2 ±33.3 mlimin/100 g). However, the No significant difference was found in the MFR between MBF duringdipyridamoleadministrationwas significantly NIDDM patients with essential hypertension (n 19; lowerinNIDDM patientswithmicrovascularangina(126 ± 2.02 ±0.83) and those without essential hypertension (n 42.7 mL/min/l00 g) than that in both NIDDM patients with 18; 2.46 ±0.85). The baseline SBP in NIDDM patients with CAD (210 ±70.1 mL/min/lOO g; P < 0.01) and control hypertension (150 ±21.8 mm Hg) was significantly higher subjects (293 ± 159 mL/min/l00 g; P < 0.01). The MBF than that in NIDDM patients without hypertension (127 ± during dipyndamole administration in those with NIDDM 13.5 mm Hg; P < 0.05). The baseline RPP in hypertensive with CAD was also significantly lower compared with that NIDDM patients (9606 ±2226) was significantly higher in control subjects (P < 0.05). than that in normotensive NIDDM patients (8082 ± 1520; The MFR was significantly lower in all NIDDM patients P < 0.05). The SBP during dipyndamole administration was (2.24 ±0.83) than that in control subjects (3.69 ±1.09; P < also significantly higher in those hypertensive NIDDM 0.01).TheMFR inpatientswithNIDDM andmicrovascular patients(140 ±18.9 mm Hg) thanthat in normotensive angina (1.90 ±0.73) was significantly lower than that in subjects (123 ±20.5 mm Hg; P < 0.05). However, the RPP both control subjects (P < 0.01) and patients with NIDDM during dipyndamole administration in hypertensive NIDDM and CAD (2.59 ±0.8 1; P < 0.05). The MFR in patients with patients (10,524 ±1248) tended to be higher than that of NIDDM and CAD was also significantlylower thanthatin those normotensive NIDDM patients (9399 ±2041), but the control subjects (P < 0.01). There were no significant difference was statistically insignificant. The MFR in dia differences in average FBS, average HbAic, baseline MBF, betic patients without hypertension was significantly re and blood pressure between the 2 patient subgroups (Tables duced compared with that in control subjects (P < 0.01), as 2 and 3). However, the mean age of patients with NIDDM and was that in hypertensive diabetic patients (P < 0.05). CAD (65.3 ±6.33 y) was significantly higher than that of patientswith microvascularangina (58.9 ±8.57; P < 0.05). MFR and Insulin Resistance The MFR in men with NIDDM (n = 28; 2.07 ±0.60) was No significant difference was found in the MFR between significantly lower than that in women with NIDDM (n —9; patientswithsevereinsulinresistance(GDR, <5 mg/mm/kg 2.76 ± 1.22; P < 0.05), whereas age was comparable [<26 pimoYmin/kg]; n = 17; MFR, 2.17 ±0.74) and those between the sexes, as were the baseline MBF and SBP and with mild insulin resistance (GDR, 5 mg/mm/kg [>26 DBP at rest (Table4). The MFR in men with NIDDM and @.tmol/min/kg];n = 20; MFR, 2.33 ±1.10).

TABLE 3 Comparison Between NIDDM Patients with Microvascular Angina and Those with CAD

restMBF NIDDM Hg)WithpatientsHbA1@ (%)FBS (pmol/minlkg)GDR(pmol/minlkg)At (mL/minll00g)SBP (mmHg)DBP (mm MVA ±1.1 ±2.50 ±11.0 ±18.8 ±18.9 ±9.23 13.1MVAWith CAD8.4 7.9 ±1.810.44 9.00 ±2.2429.1 25.9 ±11.578.7 81.8 ±16.2130 143 ±18.074.6 76.0 ±

= microvascularangina.

CORONARY MIcRo@GIoPAmY IN NIDDM •Yokoyama et al. 981 TABLE 4 Comparison Between Sex of NIDDM Patients with Microvascular Angina and Those with CAD

restAgeMBFSBPDBPSex(%)(pmol/min/kg)(pmol/min/kg)(mL/min/100 HbA1@FBSGDRAt Hg)(y)M8.5 g)(mm Hg)(mm 9.42F7.6 ±1.210.5 ±2.2729.5 ±12.279.5 ±21.3136 ±20.077.3 ±10.862.1 ± ±1.27.94 ±1.4622.2 ±6.7779.9 ±17.8143 ±19.572.3 ±11.862.2 ±5.03

MFR and HyperglycemIa microvascular angina is frequently associated with NIDDM, The MFR in the well-controlled patients with NTDDM whether there is a difference in the degree of coronary (n = 16; HbA1@,<8%; MFR, 2.81 ±0.95) was significantly microangiopathybetweenthe 2 remainsuncertain.The higher than that in the poorly controlled patients (n = 21; results of this study show that the MFR in angiographically HbA1@, 8%; MFR, 1.85 ±0.76; P < 0.01). Significant normal coronary arteries in NIDDM patients was highly differences in the average FBS and HbA1@were also noted relatedto glycemic control. between these 2 groups(Table5). The baseline MBF tended Possible Mechanism for Reduced MFR in to be lower in the well-controlled group than that in the Angiographicaliy Normal Coronary Arteries in NIDDM poorly controlled group, but the difference was statistically Coronary microangiopathy might be an essential factor insignificant (P = 0.0727). Insulin resistance was compa for the reduced MFR in anatomically normal coronary rablebetweenpatientswithCAD andpatientswithmicrovas arteries in NIDDM. However, an impaired blood flow cular angina, as were the SBP, age, and plasma lipid response to dipyridamole through flow-mediated vasodila fractions(Table5). tion (endothelial dysfunction) (34) or angiographically unde Relationship Between MFR and Glycemic Control, tectablebalanceddiffuse atherosclerosis,as has been shown insulin Resistance, Fasting insulin Concentration, in diabetics (35), may be minor factors responsible for the and Plasma Lipid Fractions reducedMFR inNIDDM. The MFR correlated with the average FBS for the Lipid disorders (19—21,30—33),aging (36), or male gen previous 5 y (P < 0.01 ; r = —0.61) (Fig. 1). A significant der (19) can reduce the MFR. However, the effect of such inverse correlation between the MFR and the average HbAic factors can be negated in this study; the lipid fractions and was found for the previous 5 y (P < 0.01; r = —0.54)(Fig. percentage of male subjects were comparable among the 3 2). However, no significant relationship was found between study groups, and no significant relationships were found the MFR and GDR or between the MFR and plasma between the MFR and such factors. Although the age of concentrationsoftotalcholesterol,triglycerides,LDL,HDL, patients with CAD was significantly higher than that of and age. Multivariatestepwise regression analysis showed patients with microvascular angina, reduction of the MFR that, among the factors considered (HbA1@,FBS, and was significantly greater in those with microvascular angina existence of microvascular angina), the average FBS (F = than that in those with CAD. Thus, age could not explain the I1.1) was independently related to the MFR (r = 0.838; results. P<0.05). The association of hyperinsulinemia with microvascular Other Major Complications angina among nondiabeticsalso has been reported (16), Severity scores of retinopathy, neuropathy, and nephropa suggesting that hyperinsulinemia may be related to coronary thy weresignificantlyhigherin thosepatientswith microvas microangiopathyinnondiabeticswithmicrovascularangina. cular angina than those in patients with CAD (1.2 ±0.87 However, no significant relationship was found between the versus1.7 ±0.39 [P < 0.01]; 0.36 ±0.51 versus0.087 ± MFR and the degree of insulin resistance. This result is 0.29 [P < 0.05]; and 0.64 ±1.03 versus 0.087 ±0.29 [P < consistent with that reported previously (1 7). The lack of 0.051, respectively). hyperinsulinemia may be a factor. It has been suggested that insulin resistance is highly related to coronary macroangiopa DISCUSSION thy ratherthanmicroangiopathy(5,14). Ourresults support this speculation. Factors That Reduce MFR A relationship between the MFR and the severity of Patients with NIDDM and Microvascular Angina or CAD coronary stenosis has been reported (29). However, recent In this study, only segments perfused by anatomically investigationshave suggestedthat the MFR can be reduced normal coronary arteries were studied. A reduced MFR of in a variety of coronaryrisk factors, including hyperlipid greater severity was noted in patients with NIDDM and emia (19—21,30—33)and diabetes (4—6).Because several microvascularanginathanthat in thosewith CAD. This factorscontributedto diabetic angiopathy,controversyex finding could be attributed to the fact that CAD is a ists concerning the factors that could contribute to coronary macrovascular abnormality (with the possibility of the microangiopathyin diabeticpatients.AlthoughCAD or coexistence of microangiopathy) and that the reduced MFR

982 THEJOURNALOFNUCLEARMEDICINE•Vol. 41 •No. 6 •June2000 tote +1 +1 (“Co (@0) COLO

a) 00)Coo @-0 (@o I-E +1 +1 E

(Y,c%J‘-Co C%JCO -I 00 +1 +1 ‘-0coeD 100 120 140 160 180 200 220 240 260 280 C,) a)C FBS tON. FIGURE 1. SignificantinverserelationshipbetweenMFR In —Jo 00 +1 +1 segments perfused by normal coronary arteries and average FBSconcentration(mg/dL)inNIDDM(r = —0.61; P < 0.01). z c@c@ •0 a) in microvascular angina is caused by a microvascular 2 abnormality (with a small possibility of the coexistence of C 00 0 +1 +1 0 E angiographically undetectable balanced macrovascular dif CoeD fuse atherosclerosis) (34). The percentage of male patients 0 0 with microvascularanginawas relativelyhigherthanthatin a) 0'-@ NIDDM patients with CAD. However, when male patients •0 a-I C +1 +1 were analyzed separately, the MFR was also significantly (@) (“C,, C NCsJ lower in patients with NIDDM with microvascular angina N.N. ‘ow than that in those with CAD. Therefore, the higher percent -Jo (“Co 0) age of male patients in the microvascular angina group a-i +1 +1 cannotaccountfor the differencein the MFR betweenthe 2 C') C') patientsubgroups.The presenceof significantcoronary z @0 stenosis in 1 myocardial segment did not influence the MFR a) 0) c@c@ in the anatomically normal coronary arteries in NIDDM 2 8 LOC') C u-I- patients because the MFR was significantly lower in patients 0 £0'? +1 +1 with NTDDMwith microvascularangina than that in those c%J with CAD. The reason for the difference in the MFR 9E N. between the 2 groups is speculative. In this study, there were J more patients with LVH among those with microvascular +1 +1 (@o E N-a) c%Jc%J

0C') I Cl; +1 +1 a N.OJ •0 4 Co'

Coo 2 ‘I 0'- +1 +1 I 8@

N:ai

•0 .r, 8 9 no >0 Lo(@ HbAlc @0 CC.) 0@ zo V FIGURE 2. SignificantinverserelationshIpbetweenMFR in segments perfused by normal coronary arteries and average HbA1@(%) in NIDDM (r = —0.54;P < 0.01).

CORONARY MICROANGIOPAThYIN NIDDM •Yokoyama et a!. 983 angina (8/19) than among those with CAD (2/18). This is 5. Nabser Pi Jr. Brown RE, OskarssonH, Winnifold MD. RossenID. Maximal coronary flow reserve and metabolic coronary vasodilation in patients with another possible explanation for our findings because the diabetes mellitus. Circulation. 1995;91:635-640. MFR was significantly lower in patients with LVH. Because 6. StrauerBE, MotzW, Vogi M, SchwartzkopffB. Impairedcomnary flow reservein multivariate regression analysis showed that glycemic con NIDDM: a possible role for diabetic cardiopathy in humans. Diabetes. 1997; 46(suppl 2):S119—S124. trol was an independent factor for the reduced MFR in 7. Kim JA, BerlinerJA, NatarajanJL. Evidence thatglucose increasesmonocyte patients with NIDDM, some unknown mechanisms for the binding to human aortic endothelial cells. Diabetes. 1994;43:1103—l107. reduced MFR in microvascular angina might exist. In this 8. Folsom AR, Eckfeldt JH, WeitzmanS. et al. Relation of carotid arterywall study, severity scores of major complications—that is, thickness todiabetes mellitus, fasting glucose and insulin, body size, and physical activity:Atherosclerosis Risk in Communities(ARIC)study investigators. Stroke. retinopathy,neuropathy,and nephropathy—weresignifi 1994;25:66-73. cantly higher in patients with microvascular angina than 9. Van Hoeven KH, Factor SM. A compatison of the pathological spectnim of those in patients with CAD. These results suggest that the hypertensive, diabetic, and hypettensive-diabetic heart disease. CircuLation. 1990;82:848—855. increasedsusceptibilityto hyperglycemiaof vascularcells in 10. YeginA, OzbnT,YeginH. Glycationoflipoproteinandacceleratedatherosclero several organs as well as neural cells or retinal cells in sis in non-insulin-dependent diabetes mellitus. In: I Clin Lab Res. 1995;25:157— NIDDM patients with microvascular angina could be a 161. 11. The Diabetes Control and Complications Thal Research Group. The effect of possible mechanism for the reduced MFR. intensive treatment of diabetes on the development and progression of long-term complicationsin insuhn-dependentdiabetes mellitus.NEngIJMed. 1993;329:977- 986. Glycemic Control and Coronary MicroangIopathy 12. GordonT,CastelliWEHjortlandMC.KannelWB.DawberTR.Diabetes,blood In this study, the reduced MFR in angiographically lipids, and the role of obesity in coronary heart disease risk for women: the normal coronary arteries was related to glycemic control but Framingham study. Ann Intern Med. l977;87:393—397. 13. LaaksoM.Glycemiccontrolandtheriskforcoronaryheartdiseaseinpatients not to insulin resistance or lipid fractions, suggesting that with non-insulin-dependent diabetes mellitus. Ann Intern Med. l996;124:127— coronary microangiopathy is most likely related to glycemic 130. control. These results are consistent with previous reports 14. DeFronzoRA.Insulinresistance,hyperinsulinemia,andcoronaryarterydisease:a complex metabolic web. Comn Artery Dis. 1992;3:l1—25. thatshowed a relationshipbetweenchronicglycemic control 15. Syvanne M, Kahn J, VIrtanen KS, Taskinen MR. HDLs containing apolipopro and the MFR in NIDDM patients without evidence of teins A-I, A-il (LpA-I:A-il) as markers of CAD in men with non-insulin ischemia (1 7,18). A relationship between hyperglycemia and dependent diabetes mellitus. CircuLation. 1995;92:364—370. 16. ChauhanA, Foote J, Petch M, Schofield P. Hyperinsulinemia,coronaryartery macrovascular complications has been suggested in diabetic disease and syndrome X. JAm Coil Cardiol. 1994;23:364—368. animals (37) and diabetic patients (7,8,10,12,13). Glycemic 17. YokoyamaI,ObtakeT,MomomuraS,etal.Reducedmyocardialflowreservein control and its relationship to microvascular complications patients with non-insulin dependent diabetes mellitus. J Am Coil Cardiol. 1997;30: 1472—1477. also has been suggested in insulin-dependent diabetics (11), 18. Yokoyama I, Obtake 1, Momomura S, et al. Hyperglycemia but not insulin but the influence of hyperglycemiaon coronarymicroangi resistance is related to coronary flow reserve in patients with non.insulin opathy in NIDDM has remained uncertain. Whether the dependent diabetes mellitus. Diabetes. 1998;47:119—124. 19. YokoyamaI, Ohtake1, MurakamiT, NishikawaJ, SasakiY, OmataM. Reduced reduced MFR can be attributed to coronary microangiopa coronaryflowreserve in familialhypercholesterolemia.JNuciMed. 199&37:1937- thy, coronary macroangiopathy, or both is uncertain (17,18). 1942. However, the results of this study indicate a relationship 20. YokoyamaI,Obtake1, Momomura5, NishikawaJ,SasakiY,OmataM.Reduced coronary flow reserve in hypercholesterolemic patients without overt coronary between glycemic control and microangiopathy in these stenosis. Circulation. 1996;94:3232—3238. patients. Furthermore, this relationship is more prominent in 21. Yokoyama I, Ohtake T, Momomura 5, et al. Reduced myocardial vasodilatation in NIDDM patients with microvascular angina than that in hypenglyceridemic patients without overt coronary stenosis. Arterioscier Thmmb Vasc Biol. l998;18:294—299. NIDDM patients with CAD. 22. Rozanski A, Diamond GA, Foi@ester JS. Berman DS, Moths D, Swan JC. Alternative referent standard for cardiac normality: implication for diagnostic testing. Ann Intern Med. 1984;l0I:164—171. CONCLUSION 23. Bottcher M, Czernin J, Sun KT, Phelps ME, Schelbert HR. Effect of caffeine on myocardial blood flow at rest and during pharmacological vasodilation. I Nuci Coronary angiopathy is seen more prominently in patients Med.l995;36:2016-202l. with NIDDM and microvascularangina than in those with 24. CzerninJ, SunK, BnmkenR, BottcherM, PhelpsM, SchelbertH. Effectof acute NIDDM and CAD. Glycemic control appears to play a andlong-termsmokingon myocardialblood flow andflow reserve.Circulation. 1995;91 :2891—2897. central role in coronary microangiopathy in patients with 25. Krivokapitch3, SmithGT.HuangSC, Ctal. ‘3N-ammoniamyocardialimagingat NIDDM. restandwithexercisein normalvolunteers.Circulation.1989;80:1328—1337. 26. KuhleWG, Porenta0, HuangSC, et al. Quantificationof regionalmyocardial blood flow using 3N-ammoniaand reoriented dynamic positron emission REFERENCES tomography.Circulation.1992;86:1004—1017. 27. Schelbert HR. Phelps ME. Hoffman EJ, Huang SC, Selin CE, Kuhi DE. N-13 1. DuranteW, SunaharaFA, San AK. Effect of diabetes on metabolic coronary ammonia as an indicator of myocardial blood flow. Circulation. 1981;63:l259— dilatation in the rat. Cardiovasc Res. 1989;23:40-45. 1272. 2. Lee JC. Downing SE. Coronary dynamics and myocardial metabolism in the 28. Campisi R, Czernin J, Karpman HL, Schelbert HR. Coronary vasodilatoiy diabeticnewbornIamb.AmJPhysiol. 1979;237:H118—H124. capacity and flow reserve in normal myocardium supplied by bypass grafts late 3. Downing SE, Lee JC, Weinstein EM. Coronary dilator actions of adenosine and after surgery. Am J Cardiol. l997;80:27—31. Co2inexperimentaldiabetes.AmJPhysiol.1982;243:H252—H258. 29. Gould KL, LipscombK. Effectsof coronarystenoses on coronaryflow reserve 4. NitenbergA, Valensi P. SachsR, Dalim M, Aptecar E, Anali JR. Impairmentof andresistance.Anti Cardiol.1974;34:48—55. coronary vascular reserve and Ach-induced coronary vasodilation in diabetic 30. DayanikliF,GrambowD, Muzil0, Mows L, RubenfireM, SchwaigerM. Early @entswith angiographically normal coronary azteiies and normal vennicular detection of abnormal coronary flow reserve in asymptomatic men at high risk for systolic function. Diabetes. 1993:32:1017—1023. CAD usingpositronemissiontomography.Circulation.1994;90:808—817.

984 THE JOURNALOF NUCLEARMEDICINE •Vol. 41 •No. 6 •June 2000 31 . Pitkaenen OP, Raitakari ET, Niinikoshi H, et al. Coronary flow reserve is impaired antiaggregating and vasodilator activity of nitric oxide. Eur J Pharmacol. in young men with familial hypercholesterolemia. J Am Coil Cardiol. 1996;28: 1991:199:1-8. I705—17 11. 35. Mintz GS, Painter IA, Pichard AD, et al. Atherosclerosis in angiographically 32. Raitakari OT, Pitkanen OP. Lehtimaki T, et al. In vivo low density lipoprotein normal coronary artery reference segments: an intravascular ultrasound study with oxidation relates to coronary reactivity in young men. J Am Coil Cardiol. clinical correlations. JAm Coil Cardiol. l995;25: 1479—1485. 1997;30:97—102. 36. Czemin J, Muller P. Chan 5, et al. Influence of age and hemodynamics on 33. Yokoyama I. Ohtake T, Momomura 5. et al. Reduced myocardial vasodilatation in myocardial blood flow and flow reserve. Circulation. 1993;88:62-69. response to dipyridamole in hypeririglyceridemia. J Am Coil Cardiol. 1998;31: 37. Yamazaki Y, Kawamon R, Matsushima H, et al. Asymptomatic hyperglycemia is 1568—1574. associated with increased intimal plus medial thickness of the carotid artery. 34. Bult H, Fret HRL, Jordeans FE, Herman AG. Dipyridamole potentiates the Diabetologia. 1995:38:585—591.

CORONARYMICROANGIOPAThYIN NIDDM •Yokoyama et al. 985