DTIC ADA627760: Autonomic Compensation to Simulated Hemorrhage Monitored With Heart Period Variability

0.32 MB·English

by  Defense Technical Information Center

Autonomic compensation to simulated hemorrhage monitored
with heart period variability
William H. Cooke, PhD; Caroline A. Rickards, PhD; Kathy L. Ryan, PhD; Victor A. Convertino, PhD
Objective: To test the hypothesis that components of heart and progressive increases of muscle sympathetic nerve activity.
periodvariabilitytrackautonomicfunctionduringsimulatedhem- Arterial pressures changed minimally and late. R-R interval time
orrhage in humans. domain variability measures and spectral power at the high
Design: Prospective experimental laboratory intervention. frequency(0.15–0.4Hz)decreasedprogressivelywithlowerbody
Setting: Human physiology laboratory. negative pressure (p < .001). Both R-R interval high-frequency
Subjects:Atotalof33healthy,nonsmoking,volunteersubjects powerandtimedomainvariabilitymeasurescorrelatedinversely
(23 men, ten women). with muscle sympathetic nerve activity and directly with pulse
Interventions: Progressive lower body negative pressure was pressure (all amalgamated R2 > .88, all p < .001).
applied in 5-min stages until the onset of impending cardiovas-
Conclusions:Componentsofheartperiodvariabilitytrackearly
cular collapse.
compensatory autonomic and hemodynamic responses to pro-
MeasurementsandMainResults:Theelectrocardiogram,beat-
gressivereductionincentralbloodvolume.Suchanalyses,inter-
by-beat finger arterial pressure, and muscle sympathetic nerve
pretedinconjunctionwithstandardvitalsigns,maycontributeto
activity from the peroneal nerve were recorded continuously.
earlierassessmentsofthemagnitudeofbloodvolumelossduring
Pulse pressure was calculated from the arterial pressure wave-
hemorrhage. (Crit Care Med 2008; 36:1892–1899)
formandusedasanestimateofrelativechangesofcentralblood
volume. Heart period variability was assessed in both time and KEYWORDS:centralhypovolemia;lowerbodynegativepressure;
stroke volume; pulse pressure; blood pressure; remote medical
frequency domains. Application of lower body negative pressure
monitoring
causedprogressivereductionsofR-Rintervalandpulsepressure
H
emorrhage is one of the during hemorrhage, until the onset of controlling mechanisms—in this case,
leading causes of death for cardiovascular decompensation, and autonomic neural function.
both civilian trauma pa- therefore provide little information on Frequency domain analysis of heart
tientsandbattlefieldcasual- early changes of volume status and sub- periodvariabilityhasbeenusedasatech-
ties (1). Despite this, first responders sequent identification of the need for nique to assess autonomic function in
have no tools to assess volume status prompt intervention in severely injured patients monitored in an intensive care
before definitive medical intervention. patients (2). unit (ICU), and differences in frequency
Standardvitalsignssuchasarterialpres- Importantly, compensation to severe characteristics were found to be associ-
sure, heart rate, and arterial oxygen– hemorrhage is driven by the autonomic ated with patient outcome (3, 4). Norris
hemoglobin saturation remain stable (if nervous system. Loss of blood volume etal.(5)havedemonstratedtheabilityof
somewhat high in the case of heart rate) triggers withdrawal of parasympathetic timedomainmeasuresofheartratevari-
and activation of sympathetic neural ac- ability to accurately predict death from
tivitytodefendagainstdecreasesofarte- traumatic injury up to 12 hrs before it
From the University of Texas at San Antonio, San rial pressure (2). Consequently, arterial occurs. Grogan et al. (6) have suggested,
Antonio,TX(WHC);andtheU.S.ArmyInstituteofSurgical pressure may fail to provide an early in- basedonthousandsofindividualICUpa-
Research,FortSamHouston,TX(CAR,KLR,VAC).
dication of hemorrhage severity because tientrecords,thatvolatility,aderivation
Theauthorshavenotdisclosedanypotentialcon-
itismaintainedfairlyconstantduringthe of beat-by-beat heart rate variability in
flictsofinterest.
Supported, in part, by the U.S. Department of early stages of mild to moderate hemor- which heart rate is sampled every 1 to 4
Defense,MedicalResearchandMaterielCommand. rhagethroughautonomicneuralmecha- secs, shows promise as a new vital sign
Theviewsexpressedhereinaretheprivateviews
nisms that increase peripheral vascular for ICU patient monitoring. We have
oftheauthorsandarenottobeconstruedasrepre-
resistance and heart rate. By the time shown previously, through frequency-
sentingthoseoftheU.S.DepartmentofDefenseorthe
U.S.DepartmentoftheArmy. arterialpressuresfalltodangerouslevels domain analysis of heart period (R-R in-
Addressforrequestsforreprintsto:WilliamH.Cooke, (whichcanoccurquickly),itmayalready terval) variability in prehospital trauma
PhD,UniversityofTexasatSanAntonio,Departmentof
be too late to apply effective lifesaving patients, that the high-frequency–to–
HealthandKinesiology,OneUTSACircle,SanAntonio,TX
interventions.Insteadofmonitoringreg- low-frequency ratio derived from a Fou-
78249.E-mail:[email protected]
Copyright©2008bytheSocietyofCriticalCare ulated variables (i.e., outcomes), such as rier transform (HF/LF) is higher in pa-
MedicineandLippincottWilliams&Wilkins arterialpressure,medicalpersonnelneed tients who eventually die compared with
DOI:10.1097/CCM.0b013e3181760d0c the capability to monitor responses of patientswhosurvivetraumaticinjury(7,
1892 CritCareMed2008Vol.36,No.6
Report Documentation Page Form Approved
OMB No. 0704-0188
Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it

does not display a currently valid OMB control number.
1. REPORT DATE 2. REPORT TYPE 3. DATES COVERED
01 JUN 2008 N/A -
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Autonomic compensation to simulated hemorrhage monitored with heart
5b. GRANT NUMBER
period variability
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER
Cooke W. H., Rickards C. A., Ryan K. L., Convertino V. A.,
5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION
United States Army Institute of Surgical Research, JBSA Fort Sam REPORT NUMBER
Houston, TX 78234
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)
11. SPONSOR/MONITOR’S REPORT
NUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release, distribution unlimited
13. SUPPLEMENTARY NOTES
14. ABSTRACT
15. SUBJECT TERMS
16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF
ABSTRACT OF PAGES RESPONSIBLE PERSON

a. REPORT b. ABSTRACT c. THIS PAGE UU 8
unclassified unclassified unclassified
Standard Form 298 (Rev. 8-98)
Prescribed by ANSI Std Z39-18
8). These differences are detectable at a from tibial or peroneal nerves at the (20).Centralhypovolemiawasinducedbyap-
point in time when standard vital signs knee)areoftenlostduetointerferenceof plication of LBNP to simulate, as closely as
such as mean arterial pressure, heart the microelectrode by increases of nega- possibleinhealthyhumanvolunteers,thehe-
rate, and arterial oxygen–hemoglobin tive chamber pressure. Negative pres- modynamicchallengesassociatedwithsevere
saturation are not different between pa- sures on the order of (cid:2)60 mm Hg and hemorrhage (16). Subjects were positioned
supine within an airtight chamber that was
tient groups. above induce physiologic responses sim-
sealedattheleveloftheiliaccrestbywayofa
If some component of heart period ilar to blood loss of (cid:1)1000 mL (16).
neoprene skirt. Because injured patients do
variability does in fact track autonomic Therefore, the issue of whether changes
notbreatheatafixedrate,wedidnotattempt
compensation to reduced central blood of heart period variability reflect in- to control breathing frequency, and subjects
volume(hemorrhage),itwouldbeanat- creases of MSNA during severe central were allowed to breathe spontaneously. Each
tractive candidate for use in noninvasive hypovolemia remains to be resolved be- subject underwent exposure to progressive
monitoring systems for trauma patients fore heart period variability can be con- LBNPuntilthepointofimpendingcardiovas-
and combat casualties. However, scien- sidered as a diagnostic tool for assessing cularcollapse.TheLBNPprotocolconsistedof
tistsandcliniciansaresplitastowhether severity of hemorrhage. a5-mincontrolperiod(baseline)followedby5
minsofchamberdecompressionat(cid:2)15,(cid:2)30,
heart period variability does (9, 10) or To address this issue, we measured
(cid:2)45, and (cid:2)60 mm Hg, and then additional
does not (11, 12) reflect modulation by MSNA from peroneal nerves directly, us-
incrementsof(cid:2)10mmHgevery5minsuntil
theautonomicnervoussystem.Paganiet ing the microneurography technique, in
the onset of cardiovascular collapse followed
al. (13) introduced the concept that the a cohort of healthy human volunteers
by a 10-min recovery period. Cardiovascular
ratio of LF/HF R-R interval spectral during simulated hemorrhage with collapsewasdefinedbyoneoracombination
powercalculatedfromfrequencydomain LBNP. The purpose of this study was to ofthefollowingcriteria:a)onsetofpresynco-
analysisprovidesanindicationofsympa- test the hypothesis that components of pal symptoms such as gray out, b) a precipi-
thetic vs. parasympathetic cardiac con- heart period variability track autonomic tousfallinsystolicpressureof(cid:1)15mmHgor
trol. This contention derives from obser- functionduringprogressivecentralhypo- a sudden bradycardia, c) progressive diminu-
vations that R-R interval low-frequency volemia of a magnitude estimated to be tionofsystolicpressureto(cid:4)70mmHg,and
d) voluntary subject termination due to dis-
oscillations (RRI ; 0.04–0.15 Hz) are similar to that of severe hemorrhage.
LF comfort from symptoms such as sweating,
modulated, importantly (but not exclu-
nausea,ordizziness.
sively),bysympatheticneuraltraffic(10,
MATERIALS AND METHODS Data Analysis. Data were sampled at 500
13). However, both Kingwell et al. (14)
Hz and digitized to computer (WINDAQ,
andSauletal.(15)failedtodemonstrate Subjects. A total of 33 nonsmoking sub- Dataq Instruments, Akron, OH). The last 3
significant correlations between directly jects (23 men, ten women) volunteered to minsofdatafromeachLBNPstageandthe3
measured muscle sympathetic nerve ac- participate in this study (age, 31 (cid:3) 2 yrs; minsjustbeforecardiovascularcollapsewere
tivity (MSNA) and RRI in research vol- height, 174 (cid:3) 2 cm; weight, 80 (cid:3) 3 kg). extractedforanalysis.IndividualR-wavesgen-
LF
unteers at rest. Furthermore, induced Before inclusion, all subjects underwent a erated from the ECG were marked at their
sympatheticactivationwithnitroprusside medicalhistoryandphysicalexaminationbya occurrence in time and used for subsequent
infusion sufficient to decrease arterial physician to ensure they had no previous or identification of systolic and diastolic pres-
current medical conditions that might pre- sures generated from the Finometer. Pulse
pressures by about 15 mm Hg revealed
clude their participation. To ensure female pressureswerecalculatedfromthedifference
nonsignificant correlations between
subjects were not pregnant, all potential fe- between systolic and diastolic pressure and
MSNA and RRI (15). Although this ob-
LF male subjects underwent a urine test within usedasanestimateofchangesinstrokevol-
servation suggests a lack of association
24hrsofexperimentation.Allsubjectsmain- ume (21). All analyses were performed with
between these variables, pharmacologic tained their normal sleep patterns, refrained commercially available data analysis software
reduction of arterial pressure does not fromexercise,andabstainedfromcaffeineand (WinCPRS,AbsoluteAliens,Turku,Finland).
representthesamephysiologiccondition other autonomic stimulants, including pre- Forrepresentationofheartperiodvariabil-
as the progressive reduction of central scription or nonprescription drugs, for (cid:1)24 ity in the time domain, we calculated R-R
blood volume that accompanies severe hrsbeforethestudy.Subjectsreceivedaverbal intervalSD(RRISD)andthepercentageofad-
hemorrhage. However, severe hemor- briefingandwrittendescriptionsofallproce- jacentR-Rintervalsthatvariedby(cid:1)50msecs
duresandrisksassociatedwiththestudyand (pNN50). For representation of heart period
rhage in humans cannot be studied in a
were made familiar with the laboratory, the variabilityinthefrequencydomain,R-Rinter-
controlled laboratory environment.
protocol, and procedures. Subjects were en- vals were replotted using linear interpolation
We have proposed the use of lower
couragedtoaskquestionsoftheinvestigators, andresampledat5Hz.Datathenwerepassed
body negative pressure (LBNP) as a re-
and then they signed an informed consent through a low-pass impulse response filter
search tool to simulate the progression form that had been approved by the Institu- with a cutoff frequency of 0.5 Hz. Data sets
from mild to severe hemorrhage in hu- tionalReviewBoardfortheprotectionofhu- weresubmittedtoaFouriertransformwitha
mans during the preshock state (16). It man subjects in research from Brooke Army Hanning window. The magnitude of R-R in-
has been known for some time that pro- MedicalCenterandtheU.S.ArmyInstituteof tervaloscillationswerequantifiedbycalculat-
gressive LBNP reduces heart rate vari- SurgicalResearch,FortSamHouston,TX. ing the power spectral density for the signal
ability(17,18)andincreasesMSNA(19). Experimental Protocol. Subjects were in- (total; 0.04–0.4 Hz). Signal areas were sepa-
strumentedwithastandardfour-leadelectro- rated into high-frequency (RRI ; 0.15–0.4
However, a robust assessment of the di- HF
cardiograph(ECG)torecordcardiacelectrical Hz)andlow-frequency(RRI ;0.04-0.15Hz)
rectrelationshipbetweenheartratevari- LF
potentials,andafingercufftorecordbeat-by- bands. To account for potentially large inter-
abilityandMSNAduringprogressivelyse-
beatfingerarterialpressure(FinometerBlood subject variability common with frequency
vere reductions in central blood volume
Pressure Monitor, TNO-TPD Biomedical In- analyses,wealsonormalizedRRI andRRI
(LBNP levels of (cid:1)45 mm Hg) has been strumentation,Amsterdam,TheNetherlands). to total power by dividing bothHFRRI anLdF
HF
limited because recordings of sympa- MSNA was recorded directly from peroneal RRI by the total power (the sum of RRI
LF HF
thetic traffic (recorded most commonly nerveswiththemicroneurographytechnique andRRI )(22).
LF
CritCareMed2008Vol.36,No.6 1893
80 baseline 80 -15 mm Hg 80 -30 mm Hg
u. 60 60 60
a.
A, 40 40 40
N
S
M
20 20 20
0 0 0
250 260 270 550 560 570 580 770 780 790
80 -45 mm Hg 80 -60 mm Hg 80 -70 mm Hg
u. 60 60 60
a.
A,
N 40 40 40
S
M
20 20 20
1140 1150 1160 1460 1470 1480 1740 1750 1760
time, s time, s
time, s
Figure1.Musclesympatheticnerveactivity(MSNA)fromaperonealnerveisshownforonesubjectduringprogressiveincreasesoflowerbodynegative
pressure.
Statistical Analysis. A one-way analysis of point where spontaneous bursts of nerve .011).Theprimaryphysiologicresponses
variance for repeated measures was used for trafficarenolongerobservable.Wewere and the number of subjects studied at
comparison of dependent variables across able to maintain good MSNA recordings each LBNP stage are shown in Table 1.
treatments (levels of LBNP). A two-way anal- at lower levels of negative pressure, but RRI and RRI correlated poorly
ysis of variance for repeated measures was LF HF
signal quality tended to degrade as cham- withMSNAatbaselinebeforeapplication
used for comparison of dependent variables
berpressureincreased.Forthisreasonwe of LBNP (r (cid:5) .12 for RRI and MSNA;
across gender groups (men vs. women) and LF
did not quantify total MSNA (burst area r (cid:5) .10 for RRI and MSNA; n (cid:5) 33).
treatments (levels of LBNP). Duncan’s mean HF
separationprocedurewasemployedforlevels multiplied by the number of bursts) but, ApplicationofLBNPcausedprogressivede-
of LBNP when the probability of LBNP main rather, quantified MSNA as burst rate creases of R-R interval and pulse pressure
effects was (cid:4).05. Amalgamated Pearson cor- (burstsper100heartbeats).Althoughour butchangedsystolicanddiastolicpressures
relation coefficients were calculated among protocoltookeachsubjecttocardiovascu- minimally. Systolic pressure began to de-
variables of interest. The probabilities of ob- lar collapse, we only included data from creaseat(cid:2)45mmHg,withlittlechangein
servingchanceeffectsduetoourexperimental subjects with good MSNA recordings for eitherdiastolicormeanpressure.Respira-
manipulationsarepresentedasexactpvalues
correlations with heart period variability tory rates in all 33 subjects at baseline
(23). All data are expressed as mean (cid:3) SE metrics.MSNAwasmaintainedthroughout (15 (cid:3) 1 breaths/min) did not differ statis-
unlessspecifiedotherwise.
cardiovascular collapse in 12 subjects tically(F(cid:5).532,p(cid:5).600)fromthoseatthe
(seven men and five women), and there- maximallevelofLBNP(15(cid:3)1breaths/min)
RESULTS
fore, the unequal sample sizes we report andweremaintainedatalltimeswithinthe
Average time to cardiovascular col- for each LBNP level represent primarily high-frequencyrange(.15to.4Hz).
lapse was 1816 (cid:3) 60 secs, revealing that loss of MSNA before cardiovascular col- Application of LBNP caused progres-
a majority of subjects experienced symp- lapse. Of the 12 subjects in whom nerve sive reductions of RRISD, pNN50, and
toms late into the (cid:2)60 mm Hg stage to recordingsweremaintaineduntilcardio- RRI butdidnotaffectRRI (p(cid:5).15).
HF LF
early into the (cid:2)70 mm Hg (n (cid:5) 33). No vascularcollapse,oneexperiencedsymp- pNN50 and RRI were reduced at (cid:2)30
HF
subject reported feeling anxiety or dis- toms during (cid:2)60 mm Hg, six during mm Hg (both p (cid:4) .05), whereas RRISD
comfort associated with LBNP. (cid:2)70 mm Hg, and five during (cid:2)80 mm wasreducedat(cid:2)45mmHg(p(cid:5).04)as
Application of LBNP often causes the Hg. An example of an MSNA recording compared with baseline. These patterns
microneurography electrode tip to shift maintained at high negative pressures is were not influenced by gender. Progres-
and move beyond the recording region shownforonesubjectinFigure1.MSNA sivereductionsinRRI wereobservedin
HF
within the peroneal nerve fascicle. Such increasedproportionatelywithincreasing both men (F (cid:5) 4.76, p (cid:4) .001) and
minutechangesinelectrodepositionde- levels of LBNP in both men (F (cid:5) 8.391, women (F (cid:5) 3.99, p (cid:4) .001), but RRI
LF
grade the signal in some subjects to the p (cid:4) .001) and women (F (cid:5) 3.388, p (cid:5) was unaltered across LBNP in both
1894 CritCareMed2008Vol.36,No.6
Table1. HemodynamicandsympatheticneuralresponsestoLBNP
Baseline LBNP15 LBNP30 LBNP45 LBNP60 LBNP70 LBNP80 Recovery
Variable n(cid:5)33 n(cid:5)33 n(cid:5)24 n(cid:5)19 n(cid:5)14 n(cid:5)9 n(cid:5)5 n(cid:5)10
RRI,ms 975(cid:3)23a 951(cid:3)22b 889(cid:3)28c 804(cid:3)38d 702(cid:3)47e 633(cid:3)53f 605(cid:3)55f 1020(cid:3)58a,b
HR,bpm 63(cid:3)1.4a,b 64(cid:3)1.4a,b 70(cid:3)2.6a,c 77(cid:3)3.3c 89(cid:3)5.4d 99(cid:3)7.8e 103(cid:3)8.5e 61(cid:3)3.5b
SAP,mmHg 131(cid:3)1.7a,b 132(cid:3)1.9a,b 127(cid:3)2.1b,c 122(cid:3)2.7c,d 119(cid:3)4.2d 112(cid:3)8.1e 116(cid:3)5.1d,e 136(cid:3)5.1a
DAP,mmHg 77(cid:3)1.0a 78(cid:3)1.1a,b 78(cid:3)1.1a,b 78(cid:3)1.2a,b 79(cid:3)1.8a,b,c 77(cid:3)3.8a 83(cid:3)3.0c,d 81(cid:3)2.8a,b,c
MAP,mmHg 98(cid:3)1.2a 98(cid:3)1.1a 96(cid:3)1.1a 95(cid:3)1.4a,b 94(cid:3)2.5b,c 89(cid:3)5.1d 94(cid:3)3.2e 102(cid:3)3.5a
PP,mmHg 54(cid:3)1.4a 54(cid:3)1.5a 49(cid:3)1.9b 44(cid:3)2.4c 40(cid:3)3.2d 34(cid:3)5.1e 33(cid:3)4.6e 53(cid:3)2.3a
MSNA,b/100 21(cid:3)2.0a 27(cid:3)2.3a,b 32(cid:3)2.6b,c 37(cid:3)3.5c,d 42(cid:3)3.1d 41(cid:3)3.8d 44(cid:3)5.7d 26(cid:3)5.0a,b
Valuesaremeans(cid:3)SE;RRI,R-Rinterval;HR,heartrate;SAP,systolicarterialpressure;DAP,diastolicarterialpressure;MAP,meanarterialpressure;
PP,pulsepressure;MSNA,musclesympatheticnerveactivityexpressedasthenumberofburstsper100heartbeats;LBNP,lowerbodynegativepressure.
Differencesbetweenmeanswithinarowthatsharethesameletterarenotdistinguishablestatistically(p(cid:1).05).
70 35
65 30
60 p < 0.001 25
s
m % p < 0.001
D, 55 0, 20
5
S N
RI 50 N 15
R p
45 10
40 5
35 0
0 15 30 45 60 70 80 0 0 15 30 45 60 70 80 0
2400 1600
2000
1200
1600
2 2
s s
m m p < 0.001
, F1200 , F 800
L H
RI RI
R R
800
400
400
0 0
0 15 30 45 60 70 80 0 0 15 30 45 60 70 80 0
LBNP, mm Hg LBNP, mm Hg
Figure2.Groupaveragesareshown(cid:3)SEforR-RintervalSD(RRISD),thepercentageofadjacentR-Rintervalsthatvariedby(cid:1)50msecs(pNN50),and
R-RintervaloscillationsderivedfromaFouriertransformatthelow(RRI )andhigh(RRI )frequencies.Dataidentifiedwithbracketsshowdifferences
LF HF
frombaseline.LBNP,lowerbodynegativepressure.
groups(F(cid:2)1.232,p(cid:1).324).Therewere AfternormalizationofR-Rintervalos- componentsofheartperiodvariabilitydem-
no statistically distinguishable differ- cillationstototalpower,RRI increased onstrated high, inverse correlations with
LF
ences in responses of heart period vari- and RRI decreased progressively with MSNA(R2(cid:1).95),withtheexceptionofalow
HF
ability and MSNA between men and LBNP and were significantly different correlationbetweenMSNAandRRI calcu-
LF
women(F(cid:2)1.134,p(cid:1).442).Changesof frombaselineat(cid:2)45mmHg(p(cid:4).001). latedinabsoluteunits(R2(cid:5).09).
time and frequency domain measures of Figure 3 shows associations among High amalgamated correlation coef-
heartperiodvariabilitywithLBNPforthe time and frequency domain measures ex- ficients, as shown in Figure 3, may
entire group of 33 subjects are shown in pressedinabsoluteunitsvs.MSNA.Inboth sometimesmaskahighrateoffairlylow
Figure 2. time and frequency domains, analyses of individualcorrelations.Toaddressthisissue,
CritCareMed2008Vol.36,No.6 1895
35 70
2
2 R = 0.96
R = 0.95
30 65
25 60
50, %20 D, ms 55
N S
N15 RI 50
p R
10 45
5 40
0 35
15 20 25 30 35 40 45 50 15 20 25 30 35 40 45 50
2400 1600
R2 = 0.09 R2 = 0.98
2000
1200
1600
2 2
s s
m m
, F1200 , F 800
L H
RI RI
R R
800
400
400
0 0
15 20 25 30 35 40 45 50 15 20 25 30 35 40 45 50
MSNA, b/100 hb MSNA, b/100 hb
Figure3.Groupaveragesfortimeandfrequencydomainvariables(cid:3)SEareplottedagainstmusclesympatheticnerveactivity(MSNA),withamalgamated
correlationcoefficientsshownintheupperrightcornerofeachpanel.RRISD,R-RintervalSD;RRILF,R-RintervaloscillationsderivedfromaFourier
transformatthelowfrequency,RRI ,R-RintervaloscillationsderivedfromaFouriertransformatthehighfrequency;pNN50,thepercentageofadjacent
HF
R-Rintervalsthatvariedby(cid:1)50msecs.
we chose a sample cohort of subjects who whether heart period variability provides
provided MSNA data up to (cid:2)70 mm Hg insightintoautonomicresponsestopro-
(n (cid:5) 9) and ran individual correlations gressive central hypovolemia. The pri-
Table2. Averageindividualcorrelationsamong amongprimaryassociationsofinterest.We marynewfindingsfromthisstudyare:1)
variousheartratevariabilitymetricsandMSNA chosethissamplecohortbecauseincluding time- (RRISD, pNN50) and frequency-
during LBNP from a sample of subjects from datathrough(cid:2)70mmHgprovidedamin- (RRI ) derived components of heart pe-
thestudycohortwhoprovideddataupto(cid:2)70 HF
imum number of data sets for robust cal- riod variability decrease progressively
mmHg(n(cid:5)9)
culation of individual correlation coeffi- with increases of MSNA, and 2) RRI
LF
cients.Weconsideredacorrelationof(cid:1).7 doesnotchange.Ourresultshaveimpli-
%individual
Variables MeanPearsonr r(cid:1)0.70 tobestrong.Averageindividualcorrelations cations for monitoring of bleeding pa-
withthepercentageofindividualswithcor- tients using a simple ECG and provide
RRIHF .55(cid:3).09 44% relationsof(cid:1).7areshowninTable2. insight into the utility of time and fre-
RRILF .32(cid:3).07 0% RRISD, pNN50, RRI , and normalized quencydomainanalysesofcardiacinter-
RRIHFnu .65(cid:3).08 55% HF
RRILFnu .66(cid:3).07 55% RRIHF,correlateddirectlywithdecreasesof beat intervals as markers of autonomic
RRISD .64(cid:3).07 44% pulsepressureduringLBNP(allR2(cid:1).88). and hemodynamic compensation to cen-
pNN50 .65(cid:3).06 55% NormalizedRRI correlatedinverselywith tral blood volume reductions and early
LF
pulse pressure (R2 (cid:5) .94), but absolute indicators of hemorrhage severity.
Valuesaremeans(cid:3)SE.MSNA,musclesym- RRI didnot(R2(cid:5).005). The report of the Post Resuscitative
pathetic nerve activity; RRI HF, R-R interval LF
and Initial Utility of Life Saving Efforts
highfrequencypower;RRILF,R-Rintervallow
frequency power; RRI HFnu, normalized high DISCUSSION (PULSE) trauma work group, sponsored,
inpart,bytheNationalHeart,Lung,and
frequency power, RRI LFnu, normalized low
We submitted a group of healthy hu- Blood Institute, revealed that hemor-
frequencypower;RRISD,R-Rintervalstandard
deviation; pNN50, the percentage of adjacent man volunteers to progressive reduction rhage accounts for a large proportion of
R-wavesthatvaryby50msecsormore. in central blood volume to determine deaths due to traumatic injury (1). Be-
1896 CritCareMed2008Vol.36,No.6
causeafull40%oftraumaticallyinjured periodvariabilitymayincreasethesensi- theusefulnessofRRI asatooltoassess
LF
patients die before they reach a hospital, tivityofdecision-supporttoolsforclinical patient status during hemorrhage.
one recommendation of the work group assessment of bleeding patients. However, even those investigators
was to improve or develop new ap- Assessedinthetimedomain,increases whohavechampionedtheuseofspectral
proachestomonitoringbleedingpatients ordecreasesofheartperiodvariabilityare analysis as an indicator of autonomic
(1).Currentstandardvitalsignmonitors thought to reflect increases or decreases functionconcedethatRRI ismostuse-
LF
include an ECG, periodic blood pressure of the overall magnitude of autonomic ful as an estimate of sympathetic nerve
recordings,andestimatesofarterialoxy- influences on the heart, but contribu- activity during procedures that progres-
gen–hemoglobin saturation. The need tions from either parasympathetic or sively increase sympathetic traffic (9, 33,
for a lifesaving intervention has been as- sympathetic components cannot be de- 34). Our results in the present study do
sociated with systolic blood pressure of termined (22). Nevertheless, based on not support this notion because LBNP
(cid:4)90 mm Hg (24, 25); unfortunately, by thousands of individual bedside records failed to increase RRI in even a single
LF
thetimethesepressurereductionsoccur, frompatientsmonitoredcontinuouslyin subject, despite progressive increases of
it may already be too late to effectively theICU,reductionsofheartratestandard MSNA(Figs.2and3;0%significantcor-
resuscitateableedingpatient.Cardiovas- deviations assessed in bins of 1 to 4 secs relationsbetweenRRI andMSNA;Table
LF
cular collapse can occur quickly, and are associated strongly with poor out- 2). It therefore seems likely that sympa-
withoutadvancedwarningfromstandard come (5, 6, 28). Whether poor patient theticcompensationtocentralhypovole-
vital signs. Because the autonomic ner- outcomes assessed with periodic analysis
miaasoccursduringhemorrhagecannot
voussystemisfundamentallyresponsible of heart rate standard deviations are as-
be adequately estimated or tracked di-
for maintaining stable hemodynamics, it sociated specifically with autonomic dys-
rectly with RRI . After normalization of
is appropriate to focus on autonomic re- functionhasnotbeendetermined.Inthis LF
RRI toexpressLFoscillationsasafrac-
sponses that change early. The results of regard, our experiment is unique in that LF
tionoftotalpower,normalizedRRI in-
the present investigation suggest that it provided the opportunity to assess the LF
creasesdirectlywithMSNAduringLBNP.
RRISD, pNN50, and RRI may provide relationship between heart period vari-
HF However,becausethenormalizationpro-
newly derived vital signs that could be abilityderivedbytimedomainanddirect cedure involves dividing LF by LF(cid:6)HF,
obtained from ECG waveforms available measures of increasing sympathetic
and we observed that RRI oscillations
on current medical monitors to assess nerve activity in unanesthetized humans HF
trackchangesofdirectlymeasuredMSNA
theneedforearlyinterventioninadvance undergoingprogressivereductionincen-
whereas RRI changes do not, the fact
of alterations in standard vital signs. tral blood volume. Our results support LF
thatnormalizedRRI tracksMSNAmay
Inadditiontohemorrhage,itislikely thenotionthatperiodicanalysisofheart LF
simply reflect the influence of the rela-
that injury and pain associated with rate standard deviations can be used to
tionship between RRI and MSNA.
trauma may result in highly variable assessautonomicdysfunctioninbleeding HF
Therefore, our experiment may be the
heartrateresponsestotrauma,unrelated patients and signal the need for early
first to demonstrate that RRI repre-
to blood loss per se, hence reducing the intervention. HF
sentsthemostspecificfrequency-derived
sensitivity of heart rate as a predictor of Although autonomic cardiovascular
heartperiodvariabilitymeasuretoreflect
patient outcome. This view is supported control cannot be assessed directly in
the autonomic response to hemorrhage.
by a recent analysis of 10,825 patients trauma patients (i.e., direct measures of
Autonomic responses to simulated
withbluntorpenetratingtrauma,reveal- MSNA), the results of our experiment
hemorrhageinourcohortofhealthysub-
ingthatheartratealonewasneithersen- suggest that autonomic function may be
jects were appropriate as evidenced by a
sitive nor specific in determining the estimated in casualties with mild to se-
relative lack of change in mean arterial
need for emergent intervention (26). Al- vere hemorrhage through analysis of
pressure. Compared with baseline, sys-
though heart rate had an amalgamated heart period variability in the frequency
tolic pressure was reduced at the (cid:2)45
correlation of 0.83 with MSNA during domain.Inhumans,cardiacinterbeatin-
mm Hg level, but the magnitude of
progressive LBNP in our subjects, RRI tervals oscillate at two primary frequen-
HF
change was small (about 7%), and sys-
correlatedhigher(0.98)withMSNA(Fig. cies;high(respiratory)frequencies(0.15–
3)andchangedearlier(LBNPat(cid:2)30mm 0.4 Hz) and low frequencies (0.04–0.15 tolicpressureneverfellto(cid:4)112mmHg
Hg) (Fig. 2) than heart rate (LBNP at Hz). High-frequency oscillations are on average, despite large reductions of
(cid:2)45 mm Hg) (Table 1) during central eliminated with large doses of atropine central volume as estimated from pulse
hypovolemia. Our results therefore sug- (29, 30), and low-frequency oscillations pressure(Table1).Incontrast,asearlyas
gest that an index of sympathetic nerve are reduced with propranolol (31, 32). the (cid:2)30 mm Hg level, both pNN50 and
activitymaybemoresensitiveforassess- Based largely on these autonomic block- RRIHF were reduced by about 50% com-
ment of blood loss in trauma patients adestudies,RRI oscillationshavebeen pared with baseline (Fig. 2). These data
HF
than heart rate. This hypothesis is sup- attributedprimarilytofluctuationsofva- highlight the fact that autonomic func-
ported by the observation of significant gal nerve traffic, and RRI oscillations tion changes earlier than standard vital
LF
differences in heart period variability be- havebeenattributedtoacombinationof signs, such as arterial pressure during
tween trauma patients who died com- factors,includingfluctuationsofbothva- simulated hemorrhage, and suggest that
paredwithacohortofpatientswholived, gal and sympathetic traffic (9). Although trackingmeasuresofautonomicfunction
despitetheirheartratesbeingsimilar(7, RRI oscillationsareassociatedinsome mayprovideuniqueinsightintotheclin-
LF
8, 27). Our results from the present in- way with fluctuations of sympathetic ical status of bleeding patients. Our data
vestigation are therefore consistent with neuralactivity(10,33),itisclearthatthe further indicate that the application of
those data obtained from the trauma lit- association is imperfect (14, 15). Our re- RRI asapotentialclinicaltoolforearly
HF
erature suggesting that the use of heart sults support this concept and challenge assessment of blood loss compared with
CritCareMed2008Vol.36,No.6 1897
measurementsofstandardvitalsignscan in RRI , RRISD, and pNN50 are highly thepossibilitythattraumainadditionto
HF
be applied universally to both male and correlated with pulse pressure (all R2 (cid:1) central blood loss could influence this
female patients. 0.88), we hypothesized that heart period relationship. A second potential limita-
Without substantial anxiety, pain, or variability metrics should represent a tion of the current study is the loss of
tissue injury, we are able to assess with sensitiveandspecificmarkertotrackthe nerve recordings at high levels of nega-
confidence the isolated effect of progres- effectiveness of fluid resuscitation and tivepressure.Duetoinherentintricacies
sive central hypovolemia. Our model of bloodloss.OurLBNPmodelprovidedthe associatedwithobtainingdirectmeasure-
simulated hemorrhage necessarily in- opportunity to test this hypothesis be- mentsofMSNAinhumans,coupledwith
ducesearlyphysiologiccompensationsto cause central blood volume was restored movementsofthemicroelectrodeassoci-
moderate or severe central hypovolemia, to baseline at the cessation of negative atedwithLBNP,itisdifficulttomaintain
with severe hypotension observable at pressure.Indeed,restorationofthemag- adequate nerve recordings as chamber
only the highest levels of negative pres- nitude of heart period variability (Fig. 2) pressures increase. Despite this, to our
sures and just before cardiovascular col- with concomitant restoration of pulse knowledge,theMSNAdatawepresentat
lapse. In some severely injured patients, pressure (Table 1) at the cessation of (cid:2)70 and (cid:2)80 mm Hg are the only data
relative hypotension in the prehospital LBNPsuggestshowheartperiodvariabil- availablethatprovideinsightintoassoci-
arena is sufficient to recommend imme- itymightalsoproveusefulinmonitoring ationsofheartperiodvariabilityandsym-
diate intervention (24), but by the time adequateresuscitationduringinfusionof pathetic traffic during central hypovole-
arterialpressuresfalltodangerouslevels, fluids.Thus,inadditiontomeasurements mia of a magnitude similar to severe
somepatientsmaybeunsalvageable.The of blood pressure, continuous monitor- hemorrhage.
inability to maintain adequate arterial ing of heart period variability may pro- ClinicalImplications.Monitoringand
pressurescouldberelatedimportantlyto vide medical personnel insight into the assessing circulatory volume status in
autonomic decompensation. For exam- status of central blood volume during bleeding patients from standard vital
ple, Converse et al. (35) showed that hy- bleeding and intervention. signs is problematic because vital signs
potension induced by hemodialysis was Resultsofthecurrentstudyshouldbe change relatively late during injury pro-
associated with a relative parasympa- interpreted with at least one technical gression.Monitoringofheartperiodvari-
thetic activation and sympathetic with- caveat, and that is the practical utility of ability may increase the capability to
drawal. Others have also reported para- assessing heart period variability in real identifypatientswhoareorarenotcom-
sympathetic predominance (estimated time. Due to the necessity of identifying pensating appropriately to hemorrhage.
from an elevated HF/LF) in ICU patients R-wavefiducialpointsfromtheECGwith Potential uses include remote monitor-
who eventually died of their injuries (3, absolute certainty for frequency domain inginmass-casualtysituations,prehospi-
4). In previous studies, we estimated au- analyses,andgivenanyconfoundingfac- talemergencytransport(7,8),ICUmon-
tonomicfunctionfromheartperiodvari- tors such as movement artifact, ectopic itoringofsepticpatients(36)andpatients
abilityinprehospitaltraumapatientsand beats, or electrically noisy ECG record- undergoingrenaldialysis(37)leadingpo-
foundanelevatedHF/LFinthosepatients ings,R-Rintervalfrequencydomainanal- tentially to hypotension, and other con-
whoeithereventuallydiedofhemorrhage yses may not be possible for clinical ap- ditions in which blood volume has the
(7) or from all-cause trauma (8). In all plicationsuntilfuturemonitorshavethe potential to decrease undetected. In the
these cases, an increase in parasympa- computing capability to filter, clean, and prehospital or military setting, tracking
thetic predominance estimated from an distinguish sinus node R-waves from changes of autonomic function with
elevation in RRI was associated with non-sinus node sources. On the other heart period variability from a simple
HF
increased mortality (i.e., a late indicator hand, it seems from our analyses that ECG could increase the potential to rec-
of outcome). The results from our simpletimedomainheartperiodvariabil- ognizeableedingpatientwhoseemssta-
present experiment are the first to sug- itymetricsalsotrackautonomicfunction ble but who may actually be progressing
gestthatprogressivereductionsinRRI well.Suchanalysesaremoreforgivingof to cardiovascular collapse and death.
HF
can provide an early indication of the ectopic or missing beats and may at this
sympathetic predominance associated time be the most promising approach to ACKNOWLEDGMENTS
with appropriate compensation during assessautonomicfunctionintheprehos-
reductionincentralbloodvolumebefore pital setting similar to what has been Wethankthesubjectsfortheircheer-
reaching a condition of irreversible pa- demonstrated in the ICU (5, 6, 28). ful cooperation, Mr. Gary Muniz for ex-
tient outcome. StudyLimitations.Wedonotsuggest cellenttechnicalassistance,andDr.John
Our results also confirm that pulse thatsubjectsinthepresentstudyresem- McManusforhisassistancewithphysical
pressure may provide insight into actual ble actual bleeding patients approaching examinations and medical monitoring.
volumelossbecauseittrackedthereduc- hemorrhagic shock. Subjects in our ex-
tionincentralbloodvolume.Convertino perimentwerenotanxiousorinpain,nor REFERENCES
et al. (21) have shown that arterial pulse did they suffer tissue injury, and we
pressure provides a good indication of thereforecannotspeculatehowtheaddi- 1. Carrico CJ, Holcomb JB, Chaudry IH: Post
relativereductionsinstrokevolumedur- tion of these real-world issues might af- resuscitativeandinitialutilityoflifesaving
efforts: Scientific priorities and strategic
ing central hypovolemia. More impor- fectourresults.Althoughweareunaware
planning for resuscitation research and life
tant,ourdatasuggestthatindicesofau- ofanyevidencetosuggestthatinjuryand
saving therapy following traumatic injury.
tonomic responses (e.g., RRI , RRISD, pain would alter the fundamental physi-
HF Report of the PULSE trauma work group.
andpNN50)mayprovidenoninvasivede- ologic relationship between heart period
AcadEmergMed2002,9:621–626
cision support for actual circulating variabilityandsympatheticnerveactivity 2. Orlinsky M, Shoemaker W, Reis ED, et al:
blood volume status. Because alterations observedinthisstudy,wecannotdismiss Currentcontroversiesinshockandresusci-
1898 CritCareMed2008Vol.36,No.6
tation. Surg Clin North Am 2001; 81: nephrine spillover, and muscle sympathetic terventions in trauma patients. Mil Med
1217–1262 nerve activity during human sympathetic 2005;170:7–13
3. Winchell RJ, Hoyt DB: Spectral analysis of nervous activation and failure. Circulation 26. BraselKJ,GuseC,GentilelloLM,etal:Heart
heartratevariabilityintheICU:Ameasureof 1994;90:234–240 rate:Isittrulyavitalsign?JTrauma2007;
autonomic function. J Surg Res 1996; 63: 15. SaulJP,ReaRF,EckbergDL,etal:Heartrate 62:812–817
11–16 and muscle sympathetic nerve variability 27. Batchinsky AI, Cancio LC, Salinas J, et al:
4. WinchellRJ,HoytDB:Analysisofheart-rate duringreflexchangesofautonomicactivity. PrehospitallossofR-to-Rintervalcomplexity
variability:Anoninvasivepredictorofdeath AmJPhysiol1990;258:H713–H721 is associated with mortality in trauma pa-
and poor outcome in patients with severe 16. CookeWH,RyanKL,ConvertinoVA:Lower tients.JTrauma2007;63:512–518
headinjury.JTrauma1997;43:927–933 body negative pressure as a model to study 28. Norris PR, Ozdas A, Cao H, et al: Cardiac
5. Norris PR, Morris JA, Ozdas A, et al: Heart progression to acute hemorrhagic shock in uncouplingandheartratevariabilitystratify
rate variability predicts trauma patient out- humans.JApplPhysiol2004;96:1249–1261 ICU patients by mortality: A study of 2088
comeasearlyas12h:Implicationsformil- 17. Cooke WH, Convertino VA: Heart rate vari- trauma patients. Ann Surg 2006; 243:
itary and civilian triage. J Surg Res 2005; ability and spontaneous baroreflex se- 804–814
129:122–128 quences: Implications for autonomic moni- 29. Pomeranz B, Macaulay RJB, Caudill MA, et
6. GroganEL,NorrisPR,SperoffT,etal:Vol-
toring during hemorrhage. J Trauma 2005; al:Assessmentofautonomicfunctioninhu-
atility: A new vital sign identified using a
58:798–805 mans by heart rate spectral analysis. Am J
novelbedsidemonitoringstrategy.JTrauma
18. Butler GC, Yamamoto Y, Xing HC, et al: Physiol1985;248:H151–H153
2005;58:7–14
Heart rate variability and fractal dimension 30. Fouad FM, Tarazi RC, Rerrario CM, et al:
7. Cooke WH, Salinas J, Convertino VA, et al:
duringorthostaticchallenges.JApplPhysiol Assessment of parasympathetic control of
Heartratevariabilityanditsassociationwith
1993;75:2602–2612 heart rate by a noninvasive method. Am J
mortality in pre-hospital trauma patients.
19. VictorRG,LeimbachN:Effectsoflowerbody Physiol1984;246:H838–H842
JTrauma2006;60:363–370
negative pressure on sympathetic discharge 31. SaulJP,BergerRD,AlbrechtP,etal:Trans-
8. Cooke WH, Salinas J, McManus JG, et al:
to leg muscles in humans. J Appl Physiol fer function analysis of the circulation:
Heart period variability in trauma patients
1987;63:2558–2562 Unique insights into cardiovascular regula-
maypredictmortalityandallowremotetri-
20. HagbarthKE,VallboAB:Pulseandrespira- tion.AmJPhysiol1991;261:H1231–H1245
age. Aviat Space Environ Med 2006; 77:
tory grouping of sympathetic impulses in 32. SaulJP,BergerRD,ChenMH,etal:Transfer
1107–1112
human muscle nerves. Acta Physiol Scand functionanalysisofautonomicregulation:II.
9. ParatiG,ManciaG,DiRienzoM:Point:Car-
1968;74:96–108 Respiratorysinusarrhythmia.AmJPhysiol
diovascular variability is/is not an index of
21. ConvertinoVA,CookeWH,HolcombJB:Ar- 1989;256:H153–H161
autonomiccontrolofthecirculation.JAppl
terialpulsepressureanditsassociationwith 33. MallianiA,PaganiM,LombardiF,etal:Car-
Physiol2006;101:676–678
reduced stroke volume during progressive diovascularneuralregulationexploredinthe
10. Pagani M, Montano N, Porta A, et al: Rela-
central hypovolemia. J Trauma 2006; 61: frequency domain. Circulation 1991; 84:
tionship between spectral components of
629–634 482–492
cardiovascular variabilities and direct mea-
22. TaskForceoftheEuropeanSocietyofCar- 34. MallianiA:Thepatternofsympathovagalbal-
suresofmusclesympatheticnerveactivityin
humans.Circulation1997;95:1411–1448 diology and the North American Society of anceexploredinthefrequencydomain.News
11. EckbergDL:Sympathovagalbalance:Acrit- Pacing and Electrophysiology: Heart rate PhysiolSci1999;14:111–117
ical appraisal. Circulation 1997; 96: variability:Standardsofmeasurement,phys- 35. Converse RL, Jacobsen TN, Jost CM, et al:
3224–3232 iologicalinterpretation,andclinicaluse.Cir- Paradoxical withdrawal of reflex vasocon-
12. Taylor JA, Studinger P: Counterpoint: Car- culation1996;93:1043–1065 strictionasacauseofhemodialysis-induced
diovascularvariabilityisnotanindexofau- 23. Ludwig DA: Use and misuse of p-values in hypotension. J Clin Invest 1992; 90:
tonomic control of the circulation. J Appl designedandobservationalstudies:Guidefor 1657–1665
Physiol2006;101:678–681 researchersandreviewers.AviatSpaceEnvi- 36. EllenbyMS,McNamesJ,LaiS,etal:Uncou-
13. Pagani M, Lombardi SG, Guzzetti S, et al: ronMed2005;76:675–680 pling and recoupling of autonomic regula-
Powerspectralanalysisofheartrateandar- 24. FranklinGA,BoazPW,SpainDA,etal:Pre- tion of the heart beat in pediatric septic
terial pressure variabilities as a marker of hospital hypotension as a valid indicator of shock.Shock2001;16:274–277
sympatho-vagalinteractioninmanandcon- traumateamactivation.JTrauma2000;48: 37. Cashion AK, Cowan PA, Milstead EJ, et al:
sciousdog.CircRes1986;59:178–193 1034–1039 Heartratevariability,mortality,andexercise
14. KingwellBA,ThompsonJM,KayeDM,etal: 25. HolcombJB,NilesSE,MillerCC,etal:Pre- inpatientswithend-stagerenaldisease.Prog
Heart rate spectral analysis, cardiac norepi- hospital physiologic data and life saving in- Transplant2000;10:10–16
CritCareMed2008Vol.36,No.6 1899