Cardiac Autonomic Neuropathy in Diabetes [630827]

Cardiac Autonomic Neuropathy in Diabetes
A clinical perspective
RODICA POP-BUSUI,MD, PHD
This review covers the epidemiology,
pathophysiology, clinical presenta-tion, and diagnosis of cardiac auto-
nomic neuropathy (CAN) in diabetes anddiscusses current evidence on approachesto prevention and treatment of CAN.
CASE PRESENTATION — A 26-
year-old woman with “brittle” type 1 dia-betes and severe CAN experiencedsudden cardiac death. She had a 16-yearhistory of poor diabetes control present-ing with wide blood glucose fluctuations,recurrent episodes of severe hypoglyce-mia, and hypoglycemia unawareness.Over time she developed persistent ortho-static hypotension with daily falls in sys-tolic blood pressure ranging from 30–60mmHg. These episodes had significantimpact on her daily activities and requiredintermittent therapy with the /H9251-1 agonist
midodrine. Other complications in-cluded severe gastroparesis, refractory di-arrhea, and painful diabetic peripheralneuropathy. Her last clinical examinationrevealed resting tachycardia with a fixedrate of 115 bpm, supine blood pressure of110/78 mmHg, which dropped to 70/48mmHg while standing, symmetrical ab-sent pinprick and temperature discrimi-nation in stocking distribution, and leftCharcot joint. Her last pertinent labora-tory findings were A1C 8.7%, creatinine1.9 mg/dl, microalbumin/creatinine 496mg/g, and hemoglobin 10.8 g/dl.
CAN represents a significant cause of
morbidity and mortality in diabetic pa-tients and is associated with a high risk ofcardiac arrhythmias and sudden death,possibly related to silent myocardial isch-emia. Therefore, it has important clinicaland prognostic relevance. Recent reportsof major clinical trials undermine estab-
lished thinking concerning glycemic con-trol and cardiac risk. Thus, a review of thistopic is both timely and important forphysicians to better understand how toassess the complexity of conditionspresent in patients with diabetes in orderto establish safe treatment targets.
EPIDEMIOLOGY — Diabetes af-
fects more than 23 million people in theU.S. (www.diabetes.org) and an esti-mated 250 million worldwide (www.who.int/diabetes). Diabetic neuropathies,including CAN, are a common chroniccomplication of type 1 and type 2 diabetesand confer high morbidity and mortalityto diabetic patients. The reported preva-lence of CAN varies greatly dependingon the criteria used to identify CAN andthe population studied. CAN prevalenceranges from as low as 2.5% of the primaryprevention cohort in the Diabetes Controland Complications Trial (DCCT) (1) to ashigh as 90% of patients with long-stand-ing type 1 diabetes who were potentialcandidat: [anonimizat]2). In a large cohort of patients with type1 and type 2 diabetes, Ziegler et al. (3),using predefined heart rate variability(HRV) tests and spectral analysis of theR-R intervals, found that 25.3% of pa-tients with type 1 diabetes and 34.3%
of patients with type 2 diabetes had ab-normal findings. Age, sex, and otherrisk factors may also influence CANdevelopment.
PATHOGENESIS — In diabetes,
CAN is ultimately the result of complexinteractions among degree of glycemiccontrol, disease duration, age-relatedneuronal attrition, and systolic and dia-
stolic blood pressure (4,5). Hyperglyce-mia plays the key role in the activation ofvarious biochemical pathways related tothe metabolic and/or redox state of thecell, which, in concert with impairednerve perfusion, contribute to the devel-opment and progression of diabetic neu-ropathies. Experimental data implicate anumber of pathogenic pathways that mayimpact autonomic neuronal function indiabetes including: formation of ad-vanced glycation end products, increasedoxidative/nitrosative stress with increasedfree radical production, activation of thepolyol and protein kinase C pathways, ac-tivation of polyADP ribosylation, and ac-tivation of genes involved in neuronaldamage (6–8). A detailed review of thesemechanisms and their complex interac-tions has been covered broadly (6–8) andis beyond the scope of this article.
CAN AND CARDIAC
DYSFUNCTION — Autonomic in-
nervation is the primary extrinsic controlmechanism regulating HRV and cardiacperformance. It has been shown thatchronic hyperglycemia promotes pro-gressive autonomic neural dysfunction ina manner that parallels the developmentof peripheral neuropathy, e.g., beginningdistally and progressing proximally. Thevagus nerve, the longest autonomic nerve,mediates /H1101175% of all parasympathetic
activity. Because neuropathy is seen firstin the longest fibers, the earliest mani-festations of autonomic neuropathy indiabetes tend to be associated with para-sympathetic denervation. As such, theinitial development of CAN in diabetes ischaracterized by early augmentation ofsympathetic tone (9). Our data (10) andthose of others (11) confirm that early inthe progression of CAN complicating type1 diabetes, there is a compensatory in-crease in the cardiac sympathetic tone inresponse to subclinical peripheral dener-vation. Sympathetic denervation followslater beginning at the apex of the ventri-cles and progressing toward the base(Fig. 1).
The initial augmentation in cardiac
sympathetic activity with subsequent ab-normal norepinephrine signaling and me-
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
From the Department of Internal Medicine, Division of Metabolism, Endocrinology and Diabetes, University
of Michigan, Ann Arbor, Michigan.
Corresponding author: Rodica Pop-Busui, rpbusui@umich.edu.Received 15 July 2009 and accepted 21 October 2009.DOI: 10.2337/dc09-1294© 2010 by the American Diabetes Association. Readers may use this article as long as the work is properly
cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby
marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.Reviews/Commentaries/ADA Statements
REVIEW ARTICLE
434 DIABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 care.diabetesjournals.org

tabolism, increased mitochondrial
oxidative stress (12), and calcium-dependent apoptosis (13) may contributeto myocardial injury (12,14) and may ex-plain the high risk of cardiac events andsudden death in these patients. The sym-pathetic imbalance associated with CANmay critically influence myocardial sub-strate utilization (15) and contribute tomitochondrial uncoupling (16), regionalventricular motion abnormalities, func-tional deficits, and cardiomyopathy (10).
CLINICAL SIGNS — Clinical symp-
toms of autonomic dysfunction may notappear until long after diabetes onset.However, subclinical CAN, manifested aschanges in HRV, may be detected within 1year of diagnosis in type 2 diabetes andwithin 2 years of diagnosis in type 1 dia-betes (17).
Impaired HRV
The earliest clinical indicator of CAN is adecrease in HRV. Variability in the instan-taneous beat-to-beat heart rate intervals isa function of sympathetic and parasympa-thetic activity that regulates the cardiacfunctional response to the body’s level ofmetabolic activity. In normal individualsthe heart rate has a high degree of beat-to-beat variability and HRV fluctuateswith respiration—increasing with inspi-ration and decreasing with expiration.Initially, clinical relevance of HRV wasidentified through observations that fetaldistress is preceded by alterations in beat-
to-beat intervals before any appreciablechange occurs in heart rate itself. The se-rious implications of abnormal HRV be-came apparent only in the late 1980s,when it was confirmed that HRV was astrong, independent predictor of mortal-ity after acute myocardial infarction (18).
Resting tachycardia
Resting heart rates of /H11011100 bpm with oc-
casional increments up to 130 bpm usu-ally occur later in the course of the diseaseand reflect a relative increase in the sym-pathetic tone associated with vagal im-pairment. However, increased restingheart rate may reflect other conditionssuch as anemia or thyroid dysfunctionand is not considered to provide a reliablediagnostic criterion of CAN in the absenceof other signs. A fixed heart rate that isunresponsive to moderate exercise, stress,or sleep indicates almost complete cardiacdenervation (19) and is indicative of se-vere CAN.
Exercise intolerance
Autonomic dysfunction may impair exer-cise tolerance and has been shown to re-duce heart rate, blood pressure, andcardiac output responses to exercise (19).It is generally recommended that diabeticpatients suspected to have CAN be testedwith a cardiac stress test before undertak-ing an exercise program. Patients withCAN need to rely on their perceived ex-ertion, not heart rate, to avoid hazardous
levels of exercise intensity (19).
Abnormal blood pressure regulation
At night, nondiabetic subjects exhibit apredominance of vagal tone and de-creased sympathetic tone, associated withreduction in nocturnal blood pressure(20). In diabetic CAN this pattern is al-tered, resulting in sympathetic predomi-nance during sleep and subsequentnocturnal hypertension (21). These areassociated with a higher frequency of leftventricular (LV) hypertrophy and both fa-tal and severe nonfatal cardiovascularevents in diabetic CAN subjects (22,23).
Orthostatic hypotension
In diabetes, orthostatic hypotension oc-curs largely as a consequence of efferentsympathetic vasomotor denervation,causing reduced vasoconstriction of thesplanchnic and other peripheral vascularbeds (24). Symptoms associated with or-thostatic hypotension include: lighthead-ness, weakness, faintness, dizziness, visualimpairment, and, in most severe cases, syn-cope on standing. These symptoms can beaggravated by a number of drugs including:vasodilators, diuretics, phenothiazines, in-sulin (through endothelium-dependentvasodilatation), and tricyclic antidepres-sants, a class of drugs commonly used forsymptomatic relief of pain associated withpainful diabetic neuropathy (19). As illus-trated in the case presentation, orthostatichypotension is associated with poor qual-ity of life.
CLINICAL EVALUATION
AND DIAGNOSISCRITERIA — There is no widely ac-
cepted single approach to the diagnosis ofCAN in diabetes. Assessment of HRV, or-thostatic hypotension, and 24-h bloodpressure profiles provides indexes of bothparasympathetic and sympathetic auto-nomic function and can be used in clinicalsettings. Other methods such as cardiacsympathetic imaging, microneurography,occlusion plethysmography, and barore-flex sensitivity are currently used pre-dominantly in research settings but mayfind a place in the clinical assessment ofCAN in the future.
HRV
HRV provides a noninvasive and objec-tive method for assessing cardiovagalfunction and may be derived from elec-trocardiogram recordings under pacedbreathing. Incorporating respiratory sig-
Figure 1— The autonomic innervation of the heart and the effects of diabetes. It has been shown
that in diabetes, in a fashion that parallels the development of peripheral neuropathy, which beginsat the tip of the toes and can progress proximally, neuropathy affecting the heart begins at the apexof the ventricles and progresses toward the base.Pop-Busui
care.diabetesjournals.org D IABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 435

nal analysis enables one to independently
measure both sympathetic and parasym-pathetic branches of the autonomic ner-vous system.
During the 1970s, Ewing et al. (25)
devised a number of simple bedside testsof short-term R-R differences to detectCAN in diabetic patients, including:changes in R-R with deep breathing,which measures sinus arrhythmia duringquiet respiration and primarily reflectsparasympathetic function (26); R-R re-sponse to standing, which induces reflextachycardia followed by bradycardia andis jointly vagal and baroreflex mediated;and Valsalva ratio, which evaluates car-diovagal function in response to a stan-dardized increase in intrathoracicpressure (Valsalva maneuver), primarilyparasympathetic mediated (26). Thesevalidated tests, described in detail in astatement by the American Diabetes Asso-ciation (27), are recommended for CANdiagnosis (7,26–28) and can be per-formed in the practitioner’s office. Therapid postural changes that are part ofhead-up-tilt-table testing, with/withoutpharmacological provocation, can beused for the investigation of CAN or ofpredisposition to neurally mediated (va-sovagal) syncope due to the wide range ofchanges in the autonomic input to theheart and in the R-R intervals. This testrequires specialized personnel and is notreadily available in general practice.
HRV with deep breathing is the most
widely used test of cardiovagal functionand has about /H1101180% specificity (29).
The R-R variation can be analyzed in a
number of different ways, including heartrate range, heart period range, standarddeviation (SD), E-to-I ratio (ratio of theshortest R-R during inspiration to thelongest R-R during expiration). Calcula-tion of mean circular resultant computedby vector analysis eliminates the effects oftrends in heart rate over time, attenuatingthe effect of basal heart rate and ectopicbeats (30). The Valsalva and postural testsare analyzed as the quotient of the largestand shortest R-R intervals recorded dur-ing each respective maneuver.
Normative cutpoints had been rec-
ommended originally for interpretation ofthe various HRV indexes (25). Recentstudies demonstrate that HRV is affectedmainly by age, rate of breathing, and pos-sibly sex (26,30). Therefore, adjustmentsfor these variables are recommended forhigher accuracy (19,26,30).
Cardiovagal function can also be eval-
uated using statistical indexes in the timeand frequency domains. Time domain
measures of the normal R-R intervals in-clude mean normal-to-normal (NN) in-terval, mean heart rate, the differencebetween the longest and shortest NN in-terval, and the variation during the differ-ence between night and day heart rate(18). Twenty-four–hour R-R recordingsallow calculation of more complex statis-tical time domain measures, such as SD ofall normal R-R intervals (SDNN), SD of5-min average of normal R-R intervals(SDANN), root–mean square of the dif-ference of successive R-R intervals(rMSSD), and the number of instances perhour in which two consecutive R-R inter-vals differ by /H1102250 ms over 24 h (pNN50).
SDNN is thought to represent joint sym-pathetic and parasympathetic modula-tion of HRV, and rMSSD and pNN50 arespecific for the parasympathetic limb(18). The accuracy of these measures canbe affected by various arrhythmias and re-quire normal sinus rhythm and atrioven-tricular-nodal function.
Spectral analysis of HRV is another
tool to evaluate CAN (19). It decomposesthe R-R signal into a set of sine and cosinewaves and estimates the magnitude ofvariability as a function of frequency. Themain frequency components describedare very-low-frequency components(/H110210.04 Hz) related to fluctuations in va-
somotor tone associated with thermoreg-ulation, the low-frequency component(0.04–0.15 Hz) associated with thebaroreceptor reflex, and the high-frequencycomponents (0.15–0.4 Hz) related to re-spiratory activity (19). It is generallythought that the sympathetic systemmodulates the lower-frequency HRVcomponents, whereas the parasympa-thetic system controls the high-frequencyHRV components. Different mathemati-cal methods have been used to analyzeHRV. Fourier transform is the most com-monly chosen due to algorithm simplicityand high processing speed (18). Thismethod, limited to stationary signals, isbased on the assumption of steady-stateconditions discarding any dynamics inthe power spectrum and does not allow aprecise detection of a sudden change inautonomous tone or a precise localizationof a particular event in time when exam-ining nonstationary conditions (31). Thecontinuous wavelet transform equationsperform a time-frequency decompositionof the signal yielding a time-dependentversion of the typical low- and high-frequency peaks (31,32). Commerciallyavailable software programs using thesemethods are available for assessment of
HRV (ANSAR, Philadelphia, PA, and Ho-kanson, Bellevue, WA).
Orthostatic hypotension
Orthostatic hypotension is documentedby a fall /H1102230 mmHg in systolic or /H1102210
mmHg in diastolic blood pressure in re-sponse to a postural change from supineto standing (19). There are some contro-versial aspects related to the cut-off valuefor the diagnostic fall in systolic bloodpressure, i.e., 30 mmHg (7) or 20 mmHg(24,33), despite the definition providedby an ad hoc consensus committee in1996 (24). Recent evidence suggests thatpostural hypotension has only moderateconcordance with HRV in the diagnosis ofCAN (33).
Imaging techniques for CAN
Quantitative scintigraphic assessment ofsympathetic innervation of the humanheart is possible with positron emissiontomography (PET) and either [
123I]meta-
iodobenzylguanidine (MIBG) or [11C]-
meta -hydroxy-ephedrine ([11C]HED)
(19,34). Deficits of LV [123I]MIBG and
[11C]HED retention have been identified
in type 1 and type 2 diabetic subjects (35–
37) with (35–37) and without (10) abnor-mal cardiovascular reflex testing.Metabolically stable [
11C]HED undergoes
highly specific uptake into sympathetic
nerve varicosities via norepinephrinetransporters, and quantitative [
11C]HED
retention may be assessed in 480 inde-
pendent LV regions (34). The strikingconsistency of the evolution of the patternof denervation in type 1 diabetes supportsthe reliability of [
11C]HED to monitor
changes in cardiac sympathetic nerve
populations and evaluate early anatomi-cal regional deficits of sympathetic dener-vation (10,34,35). As an example, Fig. 2shows the polar map analysis of LV[
11C]HED retention in subjects with type
1 diabetes expressed as Zscore analysis
versus control subjects (10).
Quantitative regional measurements
of myocardial /H9252-adrenoreceptor density
can also be assessed using PET and thehigh-affinity /H9252-adrenoreceptor radioli-
gand [
11C]CGP-12177 (38). However,
postsynaptic /H9252-adrenoreceptor density
was never assessed in human diabetes.
Baroreflex sensitivity
Baroreflex sensitivity (BRS) is a techniquethat evaluates the capability to reflexivelyincrease vagal activity and decrease sym-pathetic activity in response to a suddenCardiac autonomic neuropathy in diabetes
436 DIABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 care.diabetesjournals.org

increase in blood pressure. It is used in
research protocols to assess cardiac vagaland sympathetic baroreflex function andis calculated from the measurement of theheart rate–blood pressure relation afteran intravenous bolus of phenylephrine(39). The BRS was a significant indepen-dent risk predictor of cardiac mortality inthe Autonomic Tone and Reflexes AfterMyocardial Infarction (ATRAMI) study, alarge international multicenter prospec-tive study of 1,284 patients with a recentmyocardial infarction (39). It has beenshown that the analysis of spontaneousbaroreflex sequences gives results equiv-alent to the pharmacological methods,which lead to development of techniquesbased on servoplethysmomanometry thatmeasures blood pressure in the finger on abeat-to-beat basis (Finapress) (19).
Microneurography
This technique is based on recording elec-trical activity emitted by peroneal, tibial,or radial muscle sympathetic nerves andidentification of sympathetic bursts.Bursts have a characteristic shape consist-ing of a gradual rise and fall that is usuallyconstrained by the cardiac cycle and atleast twice the amplitude of random fluc-tuations (40). Recently available fullyautomated sympathetic neurogram tech-niques provide a rapid and objectivemethod that is minimally affected by sig-nal quality and preserves beat-by-beatsympathetic neurograms (40).
Assessment of symptoms
Symptoms associated with CAN includeexercise intolerance, orthostatic intoler-ance, and syncope (41). The correlationbetween symptom scores and deficits isgenerally weak in mild CAN, as thesesymptoms usually occur late in the dis-
ease process. Low et al. (41), using a val-idated self-report measure of autonomicsymptoms in a population-based study,found that autonomic symptoms werepresent more commonly in type 1 than intype 2 diabetes. At least one autonomicsymptom was reported in 83% of a largecohort of patients with type 2 diabetes(42). The risk of CAN as measured byHRV was positively associated with thenumber of reported autonomic symp-toms (42).
CLINICAL IMPLICATIONSMortality risk
CAN is associated with a high risk of car-diac arrhythmias and with sudden death.Longitudinal studies of subjects withCAN have shown 5-year mortality rates16–50% in type 1 and type 2 diabetes,with a high proportion attributed to sud-den cardiac death (25,42,43). In theEURODIAB Prospective Cohort Study of2,787 type 1 diabetic patients, CAN wasthe strongest predictor for mortality dur-ing a 7-year follow-up, exceeding the ef-fect of traditional cardiovascular riskfactors (44). The Hoorn study reportedthat the presence of diabetic CAN dou-bled the 9-year mortality risk in an elderlycohort (45). A meta-analysis of 15 studiesincluding 2,900 subjects with diabetes re-ported a pooled relative risk of mortalityof 3.45 (95% CI 2.66–4.47) in patientswith CAN, with a progressive increase inthe risk with the increase in the number ofabnormal CAN function tests (46). Thehigher predictive value of an increasednumber of CAN abnormalities was con-firmed more recently in two other cohortsof type 1 and type 2 diabetes showing thata combined abnormality in HRV and QT
index was a strong predictor of mortalityindependent of conventional risk factors(47,48).
The increased mortality risk associ-
ated with CAN has important implica-tions for diabetes management. A fearedconsequence of rigorous glycemic controlis an increased incidence of hypoglycemia(49,50). Hypoglycemia impairs hormonaland autonomic responses to subsequenthypoglycemia (51). Hypoglycemia un-awareness may promote a reducedthreshold for malignant arrhythmias andsubsequent sudden cardiac death, whichwas a possible reason for the suddendeath experienced by the patient dis-cussed in vignette. Thus, a recent studyreported that exposure to hypoglycemialeads to impaired CAN function inhealthy volunteers (52).
Silent myocardial ischemia and
diabetic cardiomyopathyIn a meta-analysis of 12 published stud-ies, Vinik et al. (7) reported a consistentassociation between CAN and the pres-ence of silent myocardial ischemia, mea-sured by exercise stress testing, with pointestimates for the prevalence rate ratiosfrom 0.85 to 15.53. In the Detection ofIschemia in Asymptomatic Diabetics(DIAD) study of 1,123 patients with type2 diabetes, CAN was a strong predictor ofsilent ischemia and subsequent cardio-vascular events (53). The association be-tween CAN and silent ischemia hasimportant implications, as reduced ap-preciation for ischemic pain impairstimely recognition of myocardial isch-emia or infarction, thereby delayingappropriate therapy. In patients with dia-betes, presence of symptoms such acuteonset dyspnea with/without coughing, se-vere fatigue, and/or acute onset of nauseaand vomiting should raise a high index ofsuspicion for an ischemic event andprompt the appropriate measures (19).
The presence of CAN was also
linked to the development of diabeticcardiomyopathy in type 1 diabetes be-cause in these patients LV dysfunctionoften precedes or occurs in the absenceof significant coronary artery disease orhypertension. We have identified dia-stolic dysfunction early in the course oftype 1 diabetes that correlated with ab-normal cardiac sympathetic imaging(10). Further studies are needed to clar-ify the complex interactions betweenCAN, silent myocardial ischemia, andcardiomyopathy in diabetes.
Figure 2— Polar maps of [11C]HED retention in normal control subjects ( left) and type 1 diabetic
patients with CAN ( right ). The color table is set to a maximum [11C]HED retention index value
of 0.09 ml blood /H18528min/H110021/H18528ml/H110021tissue. To quantify the “extent” of cardiac sympathetic denervation,
patients’ retention index data are statistically compared with our normal population database
using Zscore analysis.Pop-Busui
care.diabetesjournals.org D IABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 437

Intraoperative and perioperative
cardiovascular instabilityObservations in diabetic patients under-going general anesthesia reported that in-dividuals with CAN required vasopressorsupport more often than those withoutCAN (19). Individuals with CAN may ex-perience a greater decline in heart rate andblood pressure during induction of anes-thesia and more severe intraoperative hy-pothermia resulting in decreased drugmetabolism and impaired wound healing(19).
Stroke
A recent study in 1,458 patients with type2 diabetes reported that presence of CAN,assessed by standard HRV testing, wasone of the strongest predictors of isch-emic stroke in this cohort together withage and hypertension (54). Earlier reportsshowed similar associations (19).
THERAPEUTIC APPROACHESGlycemic control
The DCCT demonstrated that intensiveinsulin therapy for type 1 diabetes re-duced the incidence of CAN by 53% com-pared with conventional therapy (1). TheEpidemiology of Diabetes Interventionsand Complications (EDIC) study, theprospective observational study of theDCCT cohort, has shown persistent ben-eficial effects of past glucose control onmicrovascular complications despite theloss of glycemic separation (55). Recentlywe evaluated CAN in 1,226 well-characterized EDIC participants duringthe 13th and 14th year of EDIC follow-up. We found that during EDIC CAN pro-gressed substantially in both treatmentgroups, but the prevalence and incidenceof CAN remained significantly lower inthe former intensive group than in theformer conventional group, despite simi-lar levels of glycemic control in the EDICstudy (56). Treatment group differencesin the mean A1C level during the DCCTand the EDIC study explained virtually allof the beneficial effects of intensive versusconventional therapy on risk of incidentCAN, supporting the concept that inten-sive treatment of type 1 diabetes shouldbe initiated as early as is safely possible(56).
In type 2 diabetes, the effects of gly-
cemic control are less conclusive. The VACooperative Study demonstrated no dif-ference in the prevalence of autonomicneuropathy in type 2 diabetic patients af-ter 2 years of tight glycemic control com-pared with those without tight control
(57). On the other hand, the Steno-2 Trialreported that a targeted, intensive inter-vention involving glucose control andmultiple cardiovascular risk factors re-duced the prevalence of CAN amongpatients with type 2 diabetes and mi-croalbuminuria (58).
Other therapies
Data regarding the impact of lifestyle in-terventions in preventing progression ofCAN are emerging. Strictly supervised en-durance training combined with dietarychanges was associated with weight lossand improved HRV in patients with min-imal abnormalities (19). In the DiabetesPrevention Program, indexes of CAN im-proved most in the lifestyle modificationarm compared with the metformin or pla-cebo arm.
ACE inhibitors, angiotensin receptor
blockers, or aldose reductase inhibitorsappear promising but are yet to be vali-dated (19).
Orthostatic hypotension
The treatment of orthostatic hypotensionis challenging. Nonpharmacologicaltreatments include avoidance of suddenchanges in body posture to the head-upposition; avoiding medications that ag-gravate hypotension, such as tricyclic an-tidepressants and phenothiazines; eatingsmall, frequent meals to avoid postpran-dial hypotension; and avoiding activitiesthat involve straining, since increasedintra-abdominal and intra-thoracic pres-sure decrease venous return (19). Severalphysical counter maneuvers, such as legcrossing, squatting, and muscle pump-ing can help maintain blood pressureduring daily activities by inducing in-creased cardiac filling pressures andstroke volume.
PHARMACOLOGICAL
TREATMENTS
Midodrine
Midodrine, a peripheral-selective /H9251
1-
adrenoreceptor agonist is the only Food
and Drug Administration–approvedagent for the treatment of orthostatic hy-potension in doses of 2.5–10 mg threetimes/day. Several double-blind, placebo-controlled studies have documented itsefficacy in the treatment of orthostatic hy-potension (7). It does not cross the blood-brain barrier, resulting in fewer centralside effects. The main adverse effects arepiloerection, pruritis, paresthesias, uri-
nary retention, and supine hypertension.
Fludrocortisone acetate
Fludrocortisone acetate, a synthetic min-eralocorticoid with a long duration of ac-tion, induces plasma expansion and mayenhance the sensitivity of blood vessels tocirculating catecholamines (59). The ef-fects usually occur over a 1- to 2-weekperiod. Supine hypertension, hypokale-mia, and hypomagnesemia may occur.Caution must be used, particularly inpatients with congestive heart failure, toavoid fluid overload. Treatment withfludrocortisone should begin with 0.05mg at bedtime and may be titrated grad-ually to a maximum of 0.2 mg/day.Doses up to 0.3–0.4 mg used in refrac-tory cases are associated with high riskfor hypokalemia, excessive fluid reten-
tion, hypertension, and congestive heartfailure.
Erythropoietin
Erythropoietin may improve orthostatichypotension, but the mechanism of ac-tion for this pressor effect is still unre-solved. Possibilities include the increasein red cell mass and central blood volume,correction of the normochromic normo-cytic anemia that frequently accompaniessevere CAN, and direct or indirect neuro-humoral effects on the vascular wall andvascular tone regulation mediated by theinteraction between hemoglobin and thevasodilator nitric oxide (59). Erythropoi-etin is administered subcutaneously or in-travenously at doses of 25–75 units/kgthree times a week until the hematocritlevel approaches normal followed bylower maintenance doses ( /H1101125 units/kg
three times/week) (59).
Nonselective /H9252-blockers
Nonselective /H9252-blockers, particularly
those with intrinsic sympathomimetic ac-tivity, may have a limited role in the treat-ment of orthostatic hypotension (59). Thesuggested mechanism of action of theseagents is the blockade of vasodilat-ing/H9252-2 receptors allowing unopposed
/H9251-adrenoreceptor–mediated vasocon-
striction. To date there is no clear efficacyevidence in diabetic CAN.
Clonidine
Clonidine, an /H9251-2 antagonist, produces a
central sympatholytic effect and a conse-quent decrease in blood pressure. Patientswith severe CAN have little central sym-pathetic efferent activity, and the use ofCardiac autonomic neuropathy in diabetes
438 DIABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 care.diabetesjournals.org

clonidine (0.1–0.6 mg/day) could result
in an increase in venous return without asignificant increase in peripheral vascularresistance. Its use is limited by the incon-sistent hypertensive effect and seriousside effects.
Somatostatin analogs
Somatostatin analogs (25–200 /H9262g/day)
may attenuate orthostatic hypotension inpatients with CAN by inhibiting the re-lease of vasoactive gastrointestinal pep-tides, enhancing cardiac output, andincreasing forearm and splanchnic vascu-lar resistance. However, severe cases ofhypertension were reported with their usein patients with diabetic CAN (60).
Pyridostigmine bromide
Pyridostigmine bromide, a cholinesteraseinhibitor, was recently shown to amelio-rate orthostatic hypotension by enhanc-ing ganglionic transmission withoutworsening supine hypertension (61).
CONCLUSIONS — CAN is a serious
chronic complication of diabetes and anindependent predictor of cardiovasculardisease mortality. As illustrated by thecase vignette and by the evidence pre-sented in this review, CAN is associatedwith a poor prognosis and poor quality oflife. Conclusive clinical evidence fromrandomized prospective trials supports acentral role for hyperglycemia in thepathogenesis of CAN, although othermetabolic and vascular factors contributeto the disease state. The clinical presenta-tion of CAN comprises a broad constel-lation of symptoms and deficits. Assess-ment of HRV is an easily available tool todocument the presence of CAN. Cardiacscintigraphic imaging with sympatheticanalogs offers more sensitive diagnosticalternatives for research use. The treat-ment of CAN is challenging. Recent clin-ical evidence continues to prove thebenefits of glycemic control, while thebenefits of lifestyle interventions areemerging.
Acknowledgments — R.P.B. is supported by
American Diabetes Association Grant 1-08-CR-48, Juvenile Diabetes Research Founda-tion (JDRF) Grant 1-2008-1025, and JDRF forthe Study of Complications of Diabetes Grant4-200-421.
No potential conflicts of interest relevant to
this article were reported.
R.P.B. acknowledges the work of Dr. David
Raffel, Department of Radiology, University ofMichigan, for assistance in generating the po-
lar map of [
11C]HED retention.
References
1. The effect of intensive diabetes therapy on
measures of autonomic nervous systemfunction in the Diabetes Control andComplications Trial (DCCT). Diabetolo-gia 1998;41:416–423
2. Kennedy WR, Navarro X, Sutherland DE.
Neuropathy profile of diabetic patients ina pancreas transplantation program. Neu-rology 1995;45:773–780
3. Ziegler D, Dannehl K, Mu ¨ hlen H, Spu ¨ ler
M, Gries FA. Prevalence of cardiovascularautonomic dysfunction assessed by spec-tral analysis, vector analysis, and standardtests of heart rate variation and bloodpressure responses at various stages of di-abetic neuropathy. Diabet Med 1992;9:806–814
4. Stella P, Ellis D, Maser RE, Orchard
TJ. Cardiovascular autonomic neuropa-thy (expiration and inspiration ratio) intype 1 diabetes: incidence and predictors.J Diabetes Complications 2000;14:1–6
5. Witte DR, Tesfaye S, Chaturvedi N, Eaton
SE, Kempler P, Fuller JH, EURODIABProspective Complications Study Group.Risk factors for cardiac autonomic neu-ropathy in type 1 diabetes mellitus. Dia-betologia 2005;48:164–171
6. Pacher P, Liaudet L, Soriano FG, Mabley
JG, Szabo ´ E, Szabo ´ C. The role of poly-
(ADP-ribose) polymerase activation in thedevelopment of myocardial and endothe-lial dysfunction in diabetes. Diabetes2002;51:514–521
7. Vinik AI, Maser RE, Mitchell BD, Freeman
R. Diabetic autonomic neuropathy. Dia-betes Care 2003;26:1553–1579
8. Edwards JL, Vincent AM, Cheng HT,
Feldman EL. Diabetic neuropathy: mech-anisms to management. Pharmacol Ther2008;120:1–34
9. Axelrod S, Lishner M, Oz O, Bernheim J,
Ravid M. Spectral analysis of fluctuationsin heart rate: an objective evaluation ofautonomic nervous control in chronic re-nal failure. Nephron 1987;45:202–206
10. Pop-Busui R, Kirkwood I, Schmid H, Ma-
rinescu V, Schroeder J, Larkin D, YamadaE, Raffel DM, Stevens MJ. Sympatheticdysfunction in type 1 diabetes: associa-tion with impaired myocardial blood flowreserve and diastolic dysfunction. J AmColl Cardiol 2004;44:2368–2374
11. Taskiran M, Rasmussen V, Rasmussen B,
Fritz-Hansen T, Larsson HB, Jensen GB,Hilsted J. Left ventricular dysfunction innormotensive Type 1 diabetic patients:the impact of autonomic neuropathy.Diabet Med 2004;21:524–530
12. Givertz MM, Sawyer DB, Colucci WS. An-
tioxidants and myocardial contractility:illuminating the “Dark Side” of beta-ad-renergic receptor activation? Circulation2001;103:782–783
13. Iwai-Kanai E, Hasegawa K, Araki M, Ka-
kita T, Morimoto T, Sasayama S. alpha-and beta-adrenergic pathways differen-tially regulate cell type-specific apoptosisin rat cardiac myocytes. Circulation 1999;100:305–311
14. Paulson DJ, Light KE. Elevation of serum
and ventricular norepinephrine contentin the diabetic rat. Res Commun ChemPathol Pharmacol 1981;33:559–562
15. Drake-Holland AJ, Van d V, Roemen TH,
Hynd JW, Mansaray M, Wright ZM, andNoble MI. Chronic catecholamine deple-tion switches myocardium from carbohy-drate to lipid utilisation. CardiovascDrugsTher 2001;15:111–117
16. Schrauwen P, Hoeks J, Hesselink MK. Pu-
tative function and physiological rele-vance of the mitochondrial uncouplingprotein-3: involvement in fatty acid me-tabolism? Prog Lipid Res 2006;45:17–41
17. Pfeifer MA, Weinberg CR, Cook DL, Re-
enan A, Halter JB, Ensinck JW, Porte D Jr.Autonomic neural dysfunction in recentlydiagnosed diabetic subjects Diabetes Care1984;7:447–453
18. Task Force of the European Society of Car-
diology and the North American Society ofPacing and Electrophysiology. Heart ratevariability: standards of measurement,physiological interpretation and clinical
use. Circulation 1996;93:1043–1065
19. Vinik AI, Ziegler D. Diabetic cardiovascu-
lar autonomic neuropathy. Circulation2007;115:387–397
20. Furlan R, Guzzetti S, Crivellaro W, Dassi
S, Tinelli M, Baselli G, Cerutti S, Lom-bardi F, Pagani M, Malliani A. Continuous24-hour assessment of the neural regula-tion of systemic arterial pressure and RRvariabilities in ambulant subjects. Circu-lation 1990;81:537–547
21. Spallone V, Bernardi L, Ricordi L, Solda `P ,
Maiello MR, Calciati A, Gambardella S,Fratino P, Menzinger G. Relationship be-tween the circadian rhythms of bloodpressure and sympathovagal balance indiabetic autonomic neuropathy. Diabetes1993;42:1745–1752
22. Lurbe E, Redon J, Kesani A, Pascual JM,
Tacons J, Alvarez V, Batlle D. Increase innocturnal blood pressure and progressionto microalbuminuria in type 1 diabetes.N Engl J Med 2002;347:797–805
23. Schwartz PJ, La Rovere MT, Vanoli E. Au-
tonomic nervous system and sudden car-diac death. Experimental basis andclinical observations for post-myocardialinfarction risk stratification. Circulation1992;85:I77–I91
24. Consensus Committee of the American
Autonomic Society and the AmericanAcademy of Neurology. Consensus state-ment on the definition of orthostatic hy-potension, pure autonomic failure, andmultiple system atrophy. Neurology1996;46:1470
Pop-Busui
care.diabetesjournals.org D IABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 439

25. Ewing DJ, Campbell IW, Clarke BF. As-
sessment of cardiovascular effects in di-abetic autonomic neuropathy andprognostic implications. Ann InternMed 1980;92:308–311
26. Low PA, Denq JC, Opfer-Gehrking TL,
Dyck PJ, O’Brien PC, Slezak JM. Effect ofage and gender on sudomotor and cardio-vagal function and blood pressure re-sponse to tilt in normal subjects. MuscleNerve 1997;20:1561–1568
27. Boulton AJ, Vinik AI, Arezzo JC, Bril V,
Feldman EL, Freeman R, Malik RA, MaserRE, Sosenko JM, Ziegler D, American Di-abetes Association. Diabetic neuropa-thies: a statement by the AmericanDiabetes Association. Diabetes Care2005;28:956–962
28. Assessment: clinical autonomic testing re-
port of the Therapeutics and TechnologyAssessment Subcommittee of the Ameri-can Academy of Neurology. Neurology1996;46:873–880
29. England JD, Gronseth GS, Franklin G,
Carter GT, Kinsella LJ, Cohen JA, AsburyAK, Szigeti K, Lupski JR, Latov N, LewisRA, Low PA, Fisher MA, Herrmann D,Howard JF, Lauria G, Miller RG, Poly-defkis M, Sumner AJ, American Academyof Neurology, American Association ofNeuromuscular and ElectrodiagnosticMedicine, American Academy of PhysicalMedicine and Rehabilitation. Evaluationof distal symmetric polyneuropathy: therole of autonomic testing, nerve biopsy,and skin biopsy (an evidence-based re-view). Muscle Nerve 2009;39:106–115
30. Gelber DA, Pfeifer M, Dawson B, Schumer
M. Cardiovascular autonomic nervoussystem tests: determination of normativevalues and effect of confounding vari-ables. J Auton Nerv Syst 1997;62:40–44
31. Pichot V, Gaspoz JM, Molliex S, Antonia-
dis A, Busso T, Roche F, Costes F, QuintinL, Lacour JR, Barthe ´le´my JC. Wavelet
transform to quantify heart rate variabilityand to assess its instantaneous changes.J Appl Physiol 1999;86:1081–1091
32. Toledo E, Gurevitz O, Hod H, Eldar M,
Akselrod S. Wavelet analysis of instanta-neous heart rate: a study of autonomiccontrol during thrombolysis. Am JPhysiol Regul Integr Comp Physiol 2003;284:R1079–91
33. Spallone V, Morganti R, Fedele T,
D’Amato C, Maiello MR. Reappraisal ofthe diagnostic role of orthostatic hypoten-sion in diabetes. Clin Auton Res 2009;19:58–64
34. Raffel DM, Wieland DM. Assessment of
cardiac sympathetic nerve integrity withpositron emission tomography. Nucl MedBiol 2001;28:541–559
35. Stevens MJ, Raffel DM, Allman KC, Daya-
nikli F, Ficaro E, Sandford T, WielandDM, Pfeifer MA, Schwaiger M. Cardiacsympathetic dysinnervation in diabetes:implications for enhanced cardiovascularrisk. Circulation 1998;98:961–968
36. Stevens MJ, Raffel DM, Allman KC,
Schwaiger M, Wieland DM. Regressionand progression of cardiac sympatheticdysinnervation complicating diabetes: anassessment by C-11 hydroxyephedrineand positron emission tomography. Me-tabolism 1999;48:92–101
37. Schnell O, Muhr D, Weiss M, Dresel S,
Haslbeck M, Standl E. Reduced myocar-dial 123I-metaiodobenzylguanidine up-take in newly diagnosed IDDM patients.Diabetes 1996;45:801–805
38. Caldwell JH, Link JM, Levy WC, Poole JE,
Stratton JR. Evidence for Pre- to Postsyn-aptic Mismatch of the Cardiac Sympa-thetic Nervous System in IschemicCongestive Heart Failure. J Nucl Med2008;49:234–241
39. La Rovere MT, Bigger JT, Jr, Marcus FI,
Mortara A, Schwartz PJ. Baroreflex sensi-tivity and heart-rate variability in predic-tion of total cardiac mortality aftermyocardial infarction. ATRAMI (Auto-nomic Tone and Reflexes After Myocar-dial Infarction) Investigators. Lancet1998;351:478–484
40. Hamner JW, Taylor JA. Automated quan-
tification of sympathetic beat-by-beat ac-tivity, independent of signal quality.J Appl Physiol 2001;91:1199–206
41. Low PA, Benrud-Larson LM, Sletten DM,
Opfer-Gehrking TL, Weigand SD, O’BrienPC, Suarez GA, Dyck PJ. Autonomic symp-toms and diabetic neuropathy: a popula-tion-based study. Diabetes Care 2004;27:2942–2947
42. Navarro X, Kennedy WR, and Sutherland
DE Autonomic neuropathy and survivalin diabetes mellitus: effects of pancreastransplantation. Diabetologia 1991;34(Suppl.1):S108–S112
43. O’Brien IA, McFadden JP, Corrall RJ. The
influence of autonomic neuropathy onmortality in insulin-dependent diabetes.Q J Med 1991;79:495–502
44. Soedamah-Muthu SS, Chaturvedi N,
Witte DR, Stevens LK, Porta M, Fuller JH,EURODIAB Prospective ComplicationsStudy Group. Relationship between riskfactors and mortality in type 1 diabeticpatients in Europe: the EURODIAB Pro-spective Complications Study (PCS). Di-abetes Care 2008;31:1360–1366
45. Gerritsen J, Dekker JM, TenVoorde BJ,
Kostense PJ, Heine RJ, Bouter LM,Heethaar RM, Stehouwer CD. Impairedautonomic function is associated with in-creased mortality, especially in subjectswith diabetes, hypertension, or a historyof cardiovascular disease: the HoornStudy. Diabetes Care 2001;24:1793–1798
46. Maser RE, Mitchell BD, Vinik AI, Freeman
R. The association between cardiovascu-lar autonomic neuropathy and mortalityin individuals with diabetes: a meta-anal-ysis. Diabetes Care 2003;26:1895–90147. Lykke JA, Tarnow L, Parving HH and
Hilsted J. A combined abnormality inheart rate variation and QT corrected in-terval is a strong predictor of cardiovascu-lar death in type 1 diabetes. Scand J ClinLab Invest 2008;12:1–6
48. Ziegler D, Zentai CP, Perz S, Rathmann
W, Haastert B, Do ¨ring A, Meisinger C,
KORA Study Group. Prediction of mortal-ity using measures of cardiac autonomicdysfunction in the diabetic and nondia-betic population: the MONICA/KORAAugsburg Cohort Study. Diabetes Care2008;31:556–561
49. The effect of intensive treatment of diabe-
tes on the development and progressionof long-term complications in insulin-de-pendent diabetes mellitus: the DiabetesControl and Complications Trial Re-search Group. N Engl J Med 1993;329:977–986
50. UKPDS. Intensive blood-glucose control
with sulphonylureas or insulin comparedwith conventional treatment and risk ofcomplications in patients with type 2 di-abetes (UKPDS 33). UK Prospective Dia-betes Study (UKPDS) Group. Lancet1998;352:837–853
51. Cryer PE. Hypoglycemia-associated auto-
nomic failure in diabetes. Am J Physiol En-docrinol Metab 2001;281:E1115–E1121
52. Adler GK, Bonyhay I, Failing H, Waring
E, Dotson S, Freeman R. Antecedent hy-poglycemia impairs autonomic cardiovas-cular function: implications for rigorousglycemic control. Diabetes 2009;58:360–366
53. Young LH, Wackers FJ, Chyun DA, Davey
JA, Barrett EJ, Taillefer R, Heller GV, Is-kandrian AE, Wittlin SD, Filipchuk N,Ratner RE, Inzucchi SE, DIAD Investiga-tors. Cardiac outcomes after screening forasymptomatic coronary artery disease in
patients with type 2 diabetes: the DIADstudy: a randomized controlled trial.JAMA 2009;301:1547–1555
54. Ko SH, Song KH, Park SA, Kim SR, Cha
BY, Son HY, Moon KW, Yoo KD, Park YM,Cho JH, Yoon KH, Ahn YB. Cardiovascu-lar autonomic dysfunction predicts acuteischaemic stroke in patients with Type 2diabetes mellitus: a 7-year follow-upstudy. Diabet Med 2008;25:1171–1177
55. Writing Team for the Diabetes Control
and Complications Trial/Epidemiology ofDiabetes Interventions and Complica-tions Research Group. Sustained effect ofintensive treatment of type 1 diabetesmellitus on development and progressionof diabetic nephropathy: the Epidemiol-ogy of Diabetes Interventions and Com-plications (EDIC) study. JAMA 2003;290:2159–2167
56. Pop-Busui R, Low PA, Waberski BH, Mar-
tin CL, Albers JW, Feldman EL, SommerC, Cleary PA, Lachin JM, Herman WH,DCCT/EDIC Research Group. Effects ofprior intensive insulin therapy on cardiac
Cardiac autonomic neuropathy in diabetes
440 DIABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 care.diabetesjournals.org

autonomic nervous system function in
type 1 diabetes mellitus: the DiabetesControl and Complications Trial/Epide-miology of Diabetes Interventions andComplications study (DCCT/EDIC). Cir-culation 2009;119:2886–2893
57. Azad N, Emanuele NV, Abraira C, Hen-
derson WG, Colwell J, Levin SR, NuttallFQ, Comstock JP, Sawin CT, Silbert C,Rubino FA. The effects of intensive glyce-mic control on neuropathy in the VA co-operative study on type II diabetes
mellitus (VA CSDM). J Diabetes Compli-cations 1999;13:307–313
58. Gaede P, Vedel P, Larsen N, Jensen GV,
Parving HH, Pedersen O. Multifactorialintervention and cardiovascular disease inpatients with type 2 diabetes. N EnglJ Med 2003;348:383–393
59. Freeman R. Treatment of orthostatic hy-
potension. Semin Neurol 2003;23:435–44260. Pop-Busui R, Chey W, Stevens MJ. Severe
hypertension induced by the long-actingsomatostatin analogue sandostatin LAR ina patient with diabetic autonomic neu-ropathy [In Process Citation]. J Clin En-docrinol Metab 2000;85:943–946
61. Singer W, Sandroni P, Opfer-Gehrking TL,
Suarez GA, Klein CM, Hines S, O’Brien PC,Slezak J, Low PA. Pyridostigmine treatmenttrial in neurogenic orthostatic hypoten-sion. Arch Neurol 2006;63:513–518
Pop-Busui
care.diabetesjournals.org D IABETES CARE,VOLUME 33, NUMBER 2, F EBRUARY 2010 441