O R I G I N A L P A P E R

Systolic Time Ratio by Impedance Cardiography to DistinguishPreserved vs Impaired Left Ventricular Systolic Function

in Heart Failure

L eft ventricular (LV) ejection frac-tion (EF) is the most commonly

used measure of ventricular function inpatients with heart failure (HF), andpatients are categorized based onwhether the EF is significantly belowthe normal range. The determination ofimpaired ventricular function alignstherapy with guideline-based care whilethe determination of intact ventricularfunction provokes a very different diag-nostic and therapeutic evaluation. Thus,understanding the integrity of ventri-cular function is critical in decisionmaking for HF patients. Unfortunately,frequent measurements of EF byechocardiography or radionuclide ven-triculography are not practical or cost-effective for monitoring therapy andguiding management decisions.

Impedance cardiography (ICG) is anoninvasive method for obtaining he-modynamic parameters at baseline andwith changes in electrical impedanceduring the cardiac cycle. ICG has beenshown to be accurate and reproduciblein studies comparing ICG with the ther-modilution method using a pulmonaryartery catheter.1–3 ICG also allows mea-surement of electromechanical timingintervals, such as the systolic time ratio(STR), defined as the ratio of the pre-ejection period (the time from the onsetof QRS complex on the electrocardio-gram to the opening of the aortic valve)divided by the LV ejection time (thetime from the aortic valve opening tothe aortic valve closing). Historically,systolic time intervals and ratios arewell-established measures in the evalua-tion of LV performance but are nolonger frequently used.4 When ventricu-lar systolic function is impaired, the timeto generate pressure sufficient to open

the aortic valve increases and the dura-tion of ejection decreases, resulting in ahigher value for STR.

The ability of a specific level of STRby ICG to separate HF patients withintact from patients with impaired sys-tolic function has not been studied.Thus, the purpose of this study was toevaluate the relationship between con-temporary measures of STR and EF andto determine the ability of STR to iden-tify impaired EF in patients with knownHF.

MethodsStudy Design. A retrospective chartreview was performed in patientsenrolled in a comprehensive communityHF disease management program at a

university medical center. Patients wereincluded in the analysis if they haddeterminations of EF and STR collectedwithin a 2-week interval. All patientsprovided written, informed consent.The study protocol conforms to the eth-ical guidelines of the 1975 Declarationof Helsinki as reflected in a prioriapproval by the institution’s humanresearch committee.

Ejection Fraction. The value of the EFwas determined by either multiple gatedacquisition (MUGA) ventriculographyor by 2-dimensional echocardiographyusing standard techniques by experi-enced operators. The operators wereunaware of the STR measurements.

Left ventricular ejection fraction (EF) is used to assess patients with heart failure (HF); how-ever, frequent measurements are not cost-effective. Impedance cardiography (ICG) is alow-cost, noninvasive test that measures systolic time intervals and may be a method fordetecting impaired vs intact EF. This study evaluated the relationship between EF by echo-cardiography or gated nuclear ventriculography and systolic time ratio (STR) by ICG inoutpatients with chronic HF. A retrospective chart review identified 52 patients with EFand STR measured within 2 weeks. There was an inverse correlation between STR andEF (r=)0.54; P<.001). The area under the receiver operating characteristic curve for STRto identify reduced EF was 0.86. An EF £50% and STR ‡0.50 demonstrated 93% sensi-tivity and 85% specificity. STR was able to distinguish intact (>50%) from impaired EF(£50%). STR by ICG has the potential to be a reliable method to monitor ventricular func-tion in chronic HF. Congest Heart Fail. 2008;14:261–265. �2008 Le Jacq

Brenda Thompson, RN, MS; Mark H. Drazner, MD, MSc; Daniel L. Dries, MD,MPH; Clyde W. Yancy, MDFrom the Cardiovascular Institute, University of Texas Southwestern MedicalCenter, Dallas, TX

Address for correspondence:Brenda Thompson, RN, MS, University of Texas Southwestern Medical Center,Cardiovascular Institute, 5939 Harry Hines Boulevard, Suite 603, Dallas,TX 75390-9200E-mail: bthomp@swmcdallas.orgManuscript received October 8, 2007; revised February 13, 2008;accepted March 20, 2008

doi: 10.1111/j.1751-7133.2008.00001.x

STR by ICG to distinguish LV systolic function in HF september • october 2008 261

Systolic Time Ratio. The value of theSTR was obtained from the ICGhemodynamic status report (BioZ ICGMonitor CardioDynamics, San Diego,CA). ICG uses 4 dual sensors placed onthe neck and chest to transmit a low-amplitude, high-frequency, alternatingelectrical signal to the patient’s thorax(Figure 1). Pulsatile changes in bloodvolume and velocity are measured asimpedance changes and then applied toelectrocardiographic and blood pressuremeasurements to automatically calculatehemodynamic parameters such as car-diac output, systemic vascular resistance,and the electromechanical timing inter-vals of pre-ejection period, LV ejectiontime, and STR (Figure 2).

Statistical Methods. Paired values of EFand STR were correlated using Pearson’smethod. A value of £50% for EF wasused to define impaired systolic function.Using EF £50% as the dichotomous var-iable, a receiver operator characteristiccurve was constructed and sensitivity,specificity, positive and negative predic-tive value, and accuracy were calculatedbased on the cutoff level for STR of0.50 (Analyse-it for Excel, version 1.73;Analyse-it Software, Ltd, Leeds, UnitedKingdom).

ResultsFifty-two patients met the criteria forinclusion and were included in theanalysis. Baseline characteristics of thepatients are shown in Table I. EF wasobtained using MUGA in 23 patients

Figure 1. Impedance cardiography method showing the application of dual sensors on theneck and chest.

Figure 2. Electrocardiography (ECG) and 2 impedance cardiography waveforms fromwhich fiducial points are derived. LVET indicates left ventricular ejection time; PEP,pre-ejection period.

Table I. Patient Characteristics (N=52)

VARIABLEARIABLE NO. (%)O. (%)

SexMale 35 (67)Female 17 (33)

RaceWhite 34 (65)Black 16 (31)Hispanic 2 (4)

Heart failure etiologyIschemic 14 (27)Nonischemic 38 (73)

New York Heart Association classClass I 2 (4)Class II 17 (33)Class III 2 (4)Class IV 31 (60)

STR by ICG to distinguish LV systolic function in HF september • october 2008262

(44%) and echocardiography in 29(56%). The mean time between EF andSTR measurements was 3.5±4.7 days.

The mean EF was 38%±21% (range,10%–80%). The median EF was 34%with first quartile range 0% to 20%,second quartile 20% to 34%, third quar-tile 35% to 52%, and fourth quartilerange 52% to 80%. The overall correla-tion between EF and STR was R=)0.54(P<.001). A scatter plot of values for EFand STR are shown in Figure 3.

Figure 4 shows the receiver operatingcharacteristic curve for STR in the iden-tification of EF £50%. The inflectionpoint for EF £50% coincides with STR‡0.50. The area under the curve was 0.86(95% confidence interval, 0.71–1.0).

The diagnostic properties of STR‡0.50 with EF £35%, 40%, 45%, and50% are shown in Table II. Overallaccuracy for STR ‡0.50 was 90%. Ofthe 5 patients in whom STR did notagree with the EF category, 2 had EFmeasured by MUGA EF and 3 byechocardiography.

DiscussionThe measurement of EF to quantify LVfunction is considered an importantdiagnostic step in the evaluation ofpatients with HF in order to separatethose with impaired systolic functionfrom those with intact systolic function.In this retrospective analysis of patientsin an HF clinic, STR measured by ICGdemonstrated a significant negative cor-relation with EF and was able to reason-ably distinguish EF >50% from EF£50%.

Other studies have commented onelectrocardiographic markers for identi-fying preserved LV function in HFpatients and those who are at greaterrisk for adverse outcomes. Bhatia andcolleagues5 found that patients with pre-served LV function tended to be olderwomen with diagnoses of hypertensionwith atrial fibrillation. Patients with sys-tolic HF with a high risk for negativeoutcomes were found to have lowelectrocardiographic voltage, prolongedQRS duration, and left bundle branchblock.6–8

Systolic time intervals, including thepre-ejection period, LV ejection time,

and their ratio (STR), have long beenstudied as measures of LV performanceusing phonocardiography and carotidpulse tracing analysis. Stack and col-leagues9 found a high degree of corre-lation between STR and changes incross-sectional diameter by echocardio-graphy in patients with coronary arterydisease. Using contrast ventriculogra-phy to measure EF, Garrard and

colleagues10 demonstrated a very highcorrelation of STR and EF (R=)0.90)in 68 patients with a variety of cardiacdiseases. In a study of 453 consecutivepatients without valvular heart disease,Boudoulas and colleagues11 foundagreement between normal EF andnormal STR (defined as 0.42) in 85%of cases. ICG measures of STR havebeen shown to be strongly correlated

Figure 3. Scatter plot of ejection fraction vs systolic time ratio (N=52, R=)0.54; P<.001).

Figure 4. Receiver operating characteristic curve for systolic time ratio for the identificationof ejection fraction £50%. The area under the curve is 0.86 (95% confidence interval,0.71–1.0).

STR by ICG to distinguish LV systolic function in HF september • october 2008 263

with those obtained by carotid pulsetracing, with greater precision and easeof application using ICG.12

Other ICG-measured parameters ofLV performance have been studied fortheir ability to identify reduced EF.Ramirez and colleagues13 reported thatthe combination of reduced velocityindex and reduced acceleration indexhad 70% sensitivity and 73% specificityin identifying patients with EF <50%.Their study is complementary to thecurrent study in which we looked specif-ically at STR. Moreover, they evaluatedpatients admitted to the intensive careunit with acute HF, whereas our studywas conducted in patients with chronicHF in the outpatient setting.

The role of ICG in the evaluationand management of patients with HFhas been previously reviewed.14 Morerecent, Vijayaraghavan and colleagues15

demonstrated that hemodynamic vari-ables obtained by ICG correlatedwith changes in functional status, inclu-ding various quality-of-life measures.Specifically, improvement in STR wasshown to correlate with improvement inNew York Heart Association functionalclass (r=0.70) and the 6-minute walk test

(r=0.68). In the Prospective Evaluationand Identification of Cardiac Decompen-sation by ICG Test (PREDICT) trial,an observational study of 2316 patientvisits at 21 HF centers, hemodynamicparameters obtained by ICG, includingSTR, were reasonable identifiers ofshort-term clinical events, including HFhospitalization and death.16 Castellanosand colleagues17 studied the ability ofB-type natriuretic peptide and ICGhemodynamic parameters to predict thedevelopment of a first HF event in agroup of patients referred for echocardi-ography. They reported that both B-typenatriuretic peptide and STR by ICG hadsignificant and independent predictivevalue for long-term clinical events.

The management of HF necessitatesthe occasional assessment of a patient’schanging status. Serial measurementsof EF, however, are not consideredcost-effective and are not embraced bycurrent HF guidelines. In contrast, serialdeterminations of STR are relativelyinexpensive, and changes in STR havebeen shown to correlate with changes incardiac function, such as EF. Parrott andcolleagues18 demonstrated a good corre-lation (r=)0.73) between changes in EF

and changes in STR when measuredan average of 6.6 months apart. Theauthors suggested that ICG may be a‘‘valuable and cost-effective tool’’ formonitoring patients with HF. Consis-tent with that study, our findings pro-vide additional evidence that increasedlevels of STR in chronic HF patientssuggest impaired LV function.

LimitationsOur study is limited by the relativelysmall number of patients, the lack ofsimultaneous measurement of STR andEF, and its retrospective design. The cor-relation is modest and the test appearsbetter for predicting impaired vs intactEF and not for predicting quantitativeEF measurements. A larger dataset isrequired to confirm these observations.

ConclusionsSTR measurement by the low-cost andnoninvasive method of ICG has thepotential to provide a reliable method ofdetermining impaired vs intact systolicfunction in chronic HF. Our study maysupport its role as a marker of change incardiac status.

Disclosures: Dr. Yancy was supported byresearch grants from GlaxoSmithKline,SCIOS, Inc., and NitroMed Medtronic. Hehas received speakers’ bureau honorariafrom GlaxoSmithKline, Novartis, andserves as a consultant for ARCADiscovery,SCIOS, Inc., NitroMed, Medtronic, AstraZeneca, Otsuka, and GlaxoSmithKline.Dr. Drazner has received speakers’bureau honoraria from GlaxoSmithKline.Ms. Thompson has received speakers’bureau honoraria from Medtronic, Novar-tis, Nitromed, and SCIOS, Inc.

REFERENCES

1 Van De Water JM, Miller TW. Impedancecardiography: the next vital sign technology?Chest. 2003;123(6):2028–2033.

2 Drazner M, Thompson B, Rosenberg P, et al.Comparison of impedance cardiography withinvasive hemodynamic measurements in pa-tients with heart failure secondary to ischemicor nonischemic cardiomyopathy. Am JCardiol. 2002;89(8):993–995.

3 Albert NM, Hail MD, Li J, et al. Equivalence ofthe bioimpedance and thermodilution meth-ods in measuring cardiac output in hospital-ized patients with advanced, decompensatedchronic heart failure. Am J Crit Care.2004;13:469–479.

4 Craige E. Heart sounds: phonocardiography;carotid, apexa, and jugular venous pulse

tracings; and systolic time intervals. In: Braun-wald E, ed. Heart Disease: A Textbook ofCardiovascular Medicine. 2nd ed. Philadel-phia, PA: WB Saunders Company; 1984:54–56.

5 Bhatia RS, Tu JV, Lee DS, et al. Outcome ofheart failure with preserved ejection fractionin a population-based study. N Engl J Med.2006;355:260–269.

6 Kamath SA, Meo Neto Jde P, Canham RM,et al. Low voltage on the electrocardiogram isa marker of disease severity and a risk factorfor adverse outcomes in patients with heartfailure due to systolic dysfunction. Am Heart J.2006;152(2):355–361.

7 Iuliano S, Fisher SG, Karasik PE, et al;Department of Veterans Affairs Survival Trial

of Antiarrhythmic Therapy in CongestiveHeart Failure. QRS duration and mortality inpatients with congestive heart failure. AmHeart J. 2002;143:1085–1091.

8 Baldasseroni S, Opasich C, Gorini M, et al.Left bundle-branch block is associated withincreased 1-year sudden and total mortalityrate in 5517 outpatients with congestive heartfailure: a report from the Italian network oncongestive heart failure. Am Heart J.2002;143(3):398–405.

9 StackRS,LeeCC,ReddyBP,etal.Leftventricularperformance in coronary artery disease evalu-ated with systolic time intervals and echocar-diography. Am J Cardiol. 1976;4:331–339.

10 Garrard CL, Weissler AM, Dodge HT. Therelationship of alterations in systolic time

Table II. Sensitivity, Specificity, Positive Predictive Value, Negative Predictive Value,and Accuracy of STR ‡0.50 for the Identification of Reduced Systolic Function (EF £35%,40%, 45%, and 50%)

EF, % (NO.)SENSITIVITY,

%SPECIFICITY,

%POSITIVE PREDICTIVE

VALUE, %NEGATIVE

PREDICTIVE VALUE, % ACCURACY, %

£50 (39) 93 85 95 79 90£45 (35) 91 76 89 81 67£40 (31) 90 57 76 80 60£35 (27) 96 52 68 93 52

Abbreviations: EF, ejection fraction; STR, systolic time ratio.

STR by ICG to distinguish LV systolic function in HF september • october 2008264

intervals to ejection fraction in patients withcardiac disease. Circulation. 1970;42:455–462.

11 Boudoulas H, Geleris P, Bush CA, et al.Assessment of ventricular function by com-bined noninvasive measures: factors account-ing for methodologic disparities. Int J Cardiol.1983;2:493–501.

12 Mehlsen J, Bonde J, Rehling M, et al. Acomparison of systolic time intervals measuredby impedance cardiography and carotid pulsetracing. Dan Med Bull. 1990;37:93–95.

13 Ramirez M, Marinas C, Yamamoto M, et al.Impedance cardiography in heart failurepatients in the intensive care unit: its value inthe detection of left ventricular systolic

dysfunction and correlation with the echocar-diogram. J Am Coll Cardiol. 2004;43(5suppl A):207A.

14 Yancy C, Abraham W. Noninvasive hemo-dynamic monitoring in heart failure: utiliza-tion of impedance cardiography (ICG).Congest Heart Fail. 2003;9:241–250.

15 Vijayaraghavan K, Crum S, Cherukuri S,et al. Association of impedance cardiographyparameters with changes in functional andquality-of-life measures in patients withchronic heart failure. Congest Heart Fail.2004;10(2 suppl 2):22–27.

16 Packer M, Abraham WT, Mehra MR, et al.Utility of impedance cardiography for theidentification of short-term risk of clinical

decompensation in stable patients withchronic heart failure. J Am Coll Cardiol.2006;47:2245–2252.

17 Castellanos L, Bhalla V, Isakson S, et al.B-type natriuretic peptide and impedancecardiography testing at the time of routineechocardiography predict subsequent heartfailure. J Am Coll Cardiol. 2005;11(suppl6):S123.

18 Parrott C, Burnham K, Quale C, et al.Comparison of changes in ejection fractionto changes in impedance cardiographycardiac index and systolic time ratio. Con-gest Heart Fail. 2004;10(2 suppl 2):11–13.

STR by ICG to distinguish LV systolic function in HF september • october 2008 265