HRV to Predicts Hypotension After Spinal Anesthesia

Anesthesiology 2006; 104:53745 2006 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc.

Heart Rate Variability Predicts Severe Hypotension afterSpinal AnesthesiaRobert Hanss, M.D.,* Berthold Bein, M.D.,* Hendrik Weseloh, M.S., Martin Bauer, M.D.,* Erol Cavus, M.D.,*Markus Steinfath, M.D., Jens Scholz, M.D., Peter H. Tonner, M.D.

Background: Hypotension due to vasodilatation after spinalanesthesia (SA) may be harmful. Heart rate variability, an indi-rect measure of autonomic control, may predict hypotension.Methods: One hundred patients were studied. Retrospec-

tively, heart rate variability was analyzed in 30 patients, classi-fied depending on the lowest systolic blood pressure (SBP) afterSA. Seventy patients were studied prospectively, assigned to oneof two groups by their low to high frequency ratio (LF/HF)before SA. Sensitivity and specificity of LF/HF for prediction ofdecrease of SBP greater 20% of baseline were tested.Results: Retrospective analysis showed differences of LF/HF

depending on the degree of hypotension after SA. Prospectiveanalysis demonstrated significant differences of SBP after SAdepending on baseline LF/HF (mean SD): low LF/HF (1.3 0.7) > SBP: 91 8% of baseline versus high LF/HF (5.5 2.4) > SBP: 66 10% of baseline (P < 0.05). Baseline LF/HFas well as high frequency and proportional decrease of SBPafter SA correlated significantly, in contrast to baseline hemo-dynamic parameters heart rate and SBP. A receiver operatorcurve characteristic analysis showed a sensitivity and specific-ity of LF/HF> 2.5 of 85% to predict SBP decrease of greater than20% of baseline after SA.Conclusions: Heart rate variability analysis before SA may

predict hypotension after SA with high sensitivity and specific-ity. LF/HF may be a tool to detect patients at high risk ofhypotension due to SA. This indicates that the predictive valueof LF/HF is superior to established predictors.

SPINAL anesthesia (SA) is widely used in daily clinicalroutine. Although regional anesthesia can be advanta-geous in some respects (i.e., postoperative outcome,respiratory function),1,2 hypotension after SA is a com-mon adverse event.3,4 No strategy of preventing the rela-tive hypovolemia caused by regional anesthesiaintrave-nous crystalloids and colloids as well as prophylacticintramuscular or intravenous vasopressorshas provedentirely satisfactory and applicable to all patients.5,6

Systemic hemodynamic regulation is modulated by theautonomic nervous system (ANS).7 Hypotension due tocentral neuroaxial block is mainly a result of decreasedsystemic vascular resistance after blockade of pregangli-onic sympathetic fibers. Differences in the regulation ofthe ANS among patients may explain hemodynamic dif-ferences in response to SA. Preoperative determination

of the ANS regulation may provide an opportunity todetect patients at risk of significant hemodynamic com-promise. A noninvasive method of measuring the activityof the ANS is the analysis of heart rate variability (HRV).8

Recent studies have shown the predictive value of HRVfor hypotension after SA in pregnant patients scheduledto undergo elective cesarean delivery.9,10 The sympa-thetic activity is known to be increased during pregnan-cy.11,12 Therefore, the applicability of the results onnonpregnant patients remains speculative.

We hypothesized that (1) retrospectively analyzed pre-operative HRV may differ between patients dependingon the severity of hypotension after SA and (2) prospec-tively analyzed HRV may predict hypotension after SA.

Materials and Methods

After approval of the Institutional Ethics Committee ofthe University Hospital Schleswig-Holstein, Campus Kiel,Germany, and written informed consent were obtained,100 men (American Society of Anesthesiologists physicalstatus class I or II) scheduled to undergo prostate glandbrachytherapy during SA were studied. Exclusion crite-ria were lack of sinus rhythm, chronic medication withantiarrhythmics, coexisting infections, history of bleed-ing disorder, history of diabetes, and chronic alcoholabuse. All patients received oral premedication of 3.75 or7.5 mg midazolam, depending on the clinical decision ofthe attending anesthesiologist, 30 min before anesthesia.

Heart rate variability analysis was performed accordingto the Task Force recommendations.13 Five-minute re-cordings of the fast peaks of R waves on the electrocar-diogram were detected with a sample rate of 1,024 Hz(TF4; Varia Cardio, Olomouc, Czech Republic). The beat-to-beat variability of consecutive R waves of the sinusrhythm was measured. Data were investigated based ontime as well as frequency domain analysis. For timedomain analysis, the mean interval of consecutive beat-to-beat intervals and the SD of mean beat-to-beat inter-vals were investigated, both known to reflect parasym-pathetic activity.14,15 Frequency domain analysis wasbased on fast Fourier transformation. Power spectrumdensities were calculated for low frequencies (LF: 0.040.15 Hz) and high frequencies (HF: 0.150.4 Hz) innormalized units, defined as the LF or HF proportionalpart of the total power. Breathing was controlled at arate of 1416 breaths/min as recommended for HRVmeasurements.16 Because fast Fourier transformationanalysis requires stationary data, patients were asked to

* Staff Anesthesiologist, Medical Student, Professor of Anesthesiology, Professor of Anesthesiology and Chair, Professor of Anesthesiology andVice-Chair.

Received from the Department of Anaesthesiology and Intensive Care Medi-cine, University-Hospital Schleswig-Holstein, Campus Kiel, Germany. Submittedfor publication June 23, 2005. Accepted for publication November 17, 2005.Support was provided solely from institutional and/or departmental sources.

Address correspondence to Dr. Hanss: Department of Anaesthesiology andIntensive Care Medicine, University-Hospital Schleswig-Holstein, Campus Kiel,Schwanenweg 21, 24105 Kiel, Germany. [email protected]. Individ-ual article reprints may be purchased through the Journal Web site, www.anesthesiology.org.

Anesthesiology, V 104, No 3, Mar 2006 537

Downloaded From: http://anesthesiology.pubs.asahq.org/pdfaccess.ashx?url=/data/Journals/JASA/931076/ on 05/27/2015

lie calmly in the supine position during measurements.Beat-to-beat intervals were measured and stored contin-uously on a personal computer in all patients. HRVanalysis was performed off-line by an investigatorblinded to the hemodynamic changes after SA. Artifactswere eliminated by computer-based artifact detectionfollowed by an investigators evaluation. Beats were re-jected if they varied more than 40% from the precedingbeat. These intervals were replaced by the mean of theprevious and consecutive beat-to-beat intervals. At most,5% of a specific measurement was allowed to be re-placed. Otherwise, this specific measurement was notincluded in the analysis.

MeasurementsSix events were defined for HRV analysis and measure-

ment of hemodynamic parameters (systolic blood pres-sure [SBP], heart rate [HR], and oxygen saturation):event 1: on the day before surgery (DBS); event 2: on theday of surgery (DOS), baseline before prehydration(DOS-BL), event 3: on the DOS after prehydration (PRE);event 4: 5 min after administration of SA (SA5); event5: 15 min after administration of SA (SA15), event 6 ofHRV: 45 min after administration of SA (SA45); event 6 ofhemodynamic parameters: lowest value after onset of SA.

Administration of SAAll patients received rapid infusion of 500 ml hydroxy-

ethyl starch, 6%/130/0.4 (Voluven; Fresenius Kabi, BadHomburg, Germany), as prehydration according to theliterature.17 Thereafter, standardized SA was performed:The puncture site was interlumbar space L3L4 or L4L5with the patient in a sitting position. Fifteen milligramsisobaric bupivacaine, 0.5%, was injected via a 25-gaugeSprotte needle with the side port of the needle pointingcephalad. The level of sensory blockade was aimed atT8T10. Immediately after injection, patients were po-sitioned supine again. Oxygen, 5 l/min, was adminis-tered via facemask. The level of sensory blockade wastested by pinprick test.

Group AssignmentThis study was performed in two parts: Retrospec-

tively, 30 patients were assigned to the derivation cohortto develop a risk profile based on HRV analysis beforeSA. Patients were classified into three groups dependingon the proportional decrease of baseline SBP after SA:MILD (stable hemodynamic parameters, lowest SBP 80% of baseline), MODERATE (moderate hypotension,lowest SBP 7080% of baseline), and SEVERE (severehypotension, lowest SBP 70% of baseline). HRV wasanalyzed retrospectively. The aim was to confirm ourprevious findings in a different group of patients whomay present differences in their underlying baseline ANSregulation.10 Prospectively, 70 patients were assigned tothe validation cohort. A predictive model was built to

confirm our hypothesis in another group of patients. Thecutoff point to discriminate high- and low-risk patientswas defined based on a previous study.10 It was hypoth-esized and tested prospectively that LF/HF 2.5 indi-cates a high risk of hypotension after SA, whereas LF/HF 2.5 indicates a low risk of hypotension after SA.Hypotension was defined as an SBP less than 80% ofbaseline. The number of prospective patients was de-fined according to a power analysis. It indicated that asample size of 25 patients per group would have a powerof 80% at the 5% significance level to detect a differenceof 20% of SBP. To cover for dropouts, 35 patients pergroup were analyzed. Therefore, a total of 100 patientswere enrolled in the study.

In all patients, hypotension was treated in a standard-ized manner, if necessary. Decrease of SBP within 20% ofbaseline required no treatment. Decrease of SBP to 80%to 70% of baseline required rapid infusion of another 500ml hydroxyethyl starch. If SBP increased to no treat-ment level, no further therapy was administered. If SBPremained below 80% of baseline, an intravenous vaso-pressor (0.5 ml Akrinor; AWD Pharma, Dresden, Ger-many; 0.5 ml 50 mg cafedrin-1 HCl, 2.5 mg theodrena-lin-HCl) was given. A maximum of 1 ml of thevasopressor was injected. If MODERATE patients re-ceived more than 1 ml vasopressor, they were assignedto the SEVERE group. Decrease of SBP to less than 70%of baseline was treated with intravenous vasopressorboluses of 0.5 ml and simultaneous rapid infusion of 500ml hydroxyethyl starch until SBP increased to at least80% of baseline. The total amount of vasopressor wasanalyzed for each group. HRV was not determined dur-ing vasopressor infusion. If vasopressor boluses werenecessary, HRV measurement was started thereafter.

StatisticsProspective data were analyzed using standard soft-

ware (PRISM 4.0 GraphPad Software; San Diego, CA). Alldata were checked for normal distribution using theKolmogorov-Smirnov test based on the Dallal and Wilkin-son approximation to the Lilliefors method. Normallydistributed data and normalized HRV data during differ-ent events were analyzed using two-way analysis of vari-ance factoring for event and LF/HF ratio, followed byBonferroni correction for multiple comparisons. ThePearson correlation was used to demonstrate correlationbetween LF/HF, HF, HR, and baseline SBP and degree ofhypotension (r 0.5, P 0.05). All parametric data(hemodynamic data and prospective HRV data) are ex-pressed as mean SD, and nonparametric data (retro-spective HRV data) are expressed as median, 25th75thpercentile, and range. P 0.05 was considered statisti-cally significant. Receiver operator characteristic curveanalysis was performed to evaluate sensitivity and spec-ificity of the threshold LF/HF of 2.5 as well as HF of 30%of total power to detect patients who showed a decrease

538 HANSS ET AL.

Anesthesiology, V 104, No 3, Mar 2006


of SBP of more than 20% of baseline after SA.18 Theperformance of two other parameters, baseline preoper-ative HR and SBP, was investigated by use of receiveroperator curve analysis.19

Results

DemographicsOne hundred patients completed the study according

to the protocol. All patients showed normal HR ( 100beats/min) and normal blood pressure (100 mmHg SBP 160 mmHg, 50 mmHg diastolic blood pres-sure 100 mmHg) on the DBS. Thirty patients wereassigned to one of three retrospective groups dependingon post-SA hypotension: MILD (n 13), MODERATE(n 10), and SEVERE (n 7). Seventy patients werestudied prospectively: 31 patients were assigned togroup LF/HF 2.5, and 39 patients were assigned togroup LF/HF 2.5. The groups were comparable withrespect to demographic data (age, weight, height, Amer-

ican Society of Anesthesiologists physical status), andlevel of sensory block (table 1). Preanesthetic laboratoryvalues of all patients were within normal limits. Noperioperative complications other than hypotensionwere recorded in any of the patients.

Retrospective PartHemodynamic data (table 2 and fig. 1) showed no

differences between groups at DBS and DOS-BL.Changes of SBP after SA showed the predefined differ-ences between patients (fig. 1A). The needs for vasopres-sor boluses to restore blood pressure in MODERATE andSEVERE patients were comparable: MODERATE: 0.5 0.1 ml versus SEVERE: 0.8 0.2 ml. HRV data are shownin table 3 and figure 2. Parameters LF/HF (fig. 2A) and HF(fig. 2C) differed significantly between groups at DOS-BL. Increased LF/HF decreased after prehydration inMODERATE and SEVERE patients. LF/HF of SEVERE pa-tients remained significantly higher compared with thatof MILD patients until onset of SA in SEVERE patients.Results of time domain analysis are demonstrated intable 3 as well. There were no significant differencesbetween groups found at any of the defined events.

Prospective PartHemodynamic data are shown in table 2 and figure 1.

No differences were found between groups at DBS andDOS-BL. Patients demonstrated significant differences ofSBP after SA depending on LF/HF at DOS-BL (fig. 1B). SBPof LF/HF 2.5 patients decreased to 91 8% of base-line, whereas SBP of LF/HF 2.5 patients decreasedsignificantly to 66 10% of baseline (P 0.05 vs.baseline; P 0.05 vs. LF/HF 2.5). LF/HF 2.5 patientsrequired no vasopressor intervention, whereas in LF/HF 2.5 patients, a mean of 1.0 0.1 ml was admin-istered to restore blood pressure (P 0.05). HRV data ofprospectively assigned patients are demonstrated in fig-

Table 1. Demographics, Premedication, Spinal Anesthesia

MILD MOD SEV LF/HF 2.5 LF/HF 2.5

Age, yr 71 7 68 7 68 6 70 5 69 5Weight, kg 86 10 89 13 83 11 83 10 82 9Height, cm 175 8 175 6 177 6 175 6 177 6BMI, kg/m2 28 3 28 4 27 3 27 3 26 2ASA 2.2 0.4 2.3 0.5 2.2 0.4 2.3 0.5 2.1 0.3Mid, mg 6.9 1.4 6.7 1.5 6.9 1.4 6.7 1.5 6.6 1.8LA, mg 14.4 1.0 14.7 0.8 14.7 0.8 14.5 0.5 14.5 1.0T 9 2 8 2 8 2 8 2 8 2

Data are presented as mean SD. P 0.05.

ASA American Society of Anesthesiologists physical status; BMI bodymass index; LF/HF 2.5 prospective group with low baseline low to highfrequency ratio; LF/HF 2.5 prospective group with high baseline low tohigh frequency ratio; Mid midazolam premedication; MILD retrospectivegroup with mild hypotension; MOD retrospective group with moderatehypotension; SEV retrospective group with severe hypotension; T tho-racic level of sensory block.

Table 2. Hemodynamic Parameters: Blood Pressure and Heart Rate

Event HEMO MILD MOD SEV LF/HF 2.5 LF/HF 2.5

DBS SBP, mmHg 160 22 160 24 146 28 151 18 153 21HR, beats/min 76 11 75 12 78 9 73 7 75 6

DOS-BL SBP, mmHg 143 23 144 18 141 14 145 15 146 21HR, beats/min 70 11 79 10 83 9 67 12 72 12

DOS-PRE SBP, mmHg 144 25 137 18 137 14 137 16 145 16HR, beats/min 69 14 75 11 75 10 69 12 74 11

DOS-SA15 SBP, mmHg 138 2 123 15* 123 17 131 14 121 13HR, beats/min 61 10 61 7 61 6 61 11 61 9

LOW SBP, mmHg 133 19 (93%) 116 11 (78%)* 98 5 (68%)* 125 12 (86%) 97 12 (66%)HR, beats/min 62 11 (90%) 63 8 (79%) 57 6 (68%) 61 11 (83%) 61 9 (81%)

Data are presented as mean SD.

* P 0.05 vs. MILD. P 0.05 MOD vs. SEV. P 0.05 LF/HF 2.5 vs. LF/HF 2.5. P 0.05 DOS-BL vs. LOW.

DBS day before surgery; DOS-BL day of surgery baseline; DOS-PRE day of surgery after prehydration; DOS-SA15 measurement 15 min after spinalanesthesia; HEMO hemodynamic parameter; HR heart rate; LF/HF 2.5 prospective group with low baseline low to high frequency ratio; LF/HF 2.5 prospective group high baseline low to high frequency ratio; LOW lowest values of systolic blood pressure and heart rate after spinal anesthesia, total valuesand relative decrease from baseline; MILD retrospective group with mild hypotension; MOD retrospective group with moderate hypotension; SBP systolicblood pressure; SEV retrospective group with severe hypotension.

539HEART RATE VARIABILITY PREDICTS SPINAL HYPOTENSION



ure 3. Throughout the study, LF/HF (fig. 3A) and LF (fig.3B) remained higher and HF (fig. 3C) remained lower inthe LF/HF 2.5 group compared with the LF/HF 2.5group. At DBS, HF (fig. 3C) was significantly lower in theLF/HF 2.5 group compared with the LF/HF 2.5group. Results of time domain analysis are demonstratedin table 3. There were no significant differences betweengroups found at any of the defined events.

Correlation, Sensitivity, and SpecificityCorrelations of different parameters and percentage

decrease of SBP after SA are demonstrated in figure 4.LF/HF of DOS-BL of all patients was significantly corre-lated to the percentage decrease of SBP (fig. 4A; r 0.6,P 0.001). So was the correlation of HF of DOS-BL andthe percentage decrease of SBP (fig. 4B; r 0.5, P 0.001). The established predictors HR and SBP did notcorrelate with the degree of hypotension (HR: r 0.07,P 0.5 [fig. 4C]; SBP: r 0.04, P 0.7 [fig. 4D]).Receiver operator curve analyses are demonstrated infigure 5. They revealed that baseline LF/HF (fig. 5A) issuitable to detect patients at risk, because the area underthe curve was significantly different from 0.5. In con-trast, baseline HF (fig. 5B), baseline HR (area 0.56, P0.28), and baseline SBP (area 0.53, P 0.65) are notsuitable to detect high-risk patients.

Discussion

Heart rate variability was analyzed in 100 patientsscheduled to undergo prostate gland brachytherapy dur-ing SA perioperatively. Thirty patients were assigned tothe derivation cohort to develop a risk profile based onthe degree of hypotension after SA. Seventy patients

were assigned to the validation cohort. Groups weredefined by LF/HF before SA. Retrospectively assignedpatients with severe hypotension after SA demonstrateda significantly higher LF/HF and lower HF at baseline.Prospectively assigned patients with increased LF/HF atbaseline had development of severe hypotension afterSA. LF as well as HF showed significant differences be-tween subgroup LF/HF 2.5 and subgroup LF/HF 2.5. Sympatholysis due to SA was reflected by a signifi-cant decrease of LF/HF as well as LF and a significantincrease of HF in the course of SA. Receiver operatorcurve analysis showed a high sensitivity and specificityof the parameter baseline LF/HF for prediction of hypo-tension after SA.

Invasive direct measures of autonomic control are notpossible in a clinical setting. Analysis of HRV is a nonin-vasive, thus indirect measure of autonomic regula-tion.14,15,2023 Time as well as frequency domain param-eters were investigated in this study. It has beensuggested that LF/HF may reflect the balance of the ANSregulation; LF may reflect vasomotor activity, which is anindirect index of sympathetic nerve activity as well aspartially affected by parasympathetic activity; and HF aswell as time domain parameters may reflect vagal nerveactivity.14,15,20,21 Nevertheless, interpretation of LF isless certain than interpretation of HF.24,25 Based on thestudy results and previous findings, we suggest that LFmay to some extent reflect sympathetic activity of theANS, and LF/HF may reflect the relation between sym-pathetic and parasympathetic function.10

HRV Analysis for Prediction of HypotensionRetrospective as well as prospective study design has

been introduced by our group in women scheduled to

Fig. 1. Hemodynamic data. Decrease of systolic blood pressure (SBP) demonstrated as proportional decrease from baseline (BL). (A)SBP, retrospective groups. (B) SBP, prospective groups. DOS-PRE day of surgery after prehydration; LF/HF < 2.5 baseline lowto high frequency ratio less than 2.5; LF/HF > 2.5 baseline low to high frequency ratio greater than 2.5; LOW lowest value afterspinal anesthesia; MILDmild hypotension; MODmoderate hypotension; SA5 5 min after spinal anesthesia; SA15 15 minafter spinal anesthesia; SEV severe hypotension. Data are presented as mean SD. * P < 0.05 MILD versus MOD as well as SEV. P < 0.05 versus changes within group versus baseline. # P < 0.05 LF/HF < 2.5 versus LF/HF > 2.5.

540 HANSS ET AL.



undergo elective cesarean delivery during SA.10 Retro-spective findings indicate that patients who had devel-opment of hypotension after SA had higher LF/HF andlower HF preoperatively. Results were confirmed in alarger prospective group. These results are in goodagreement with previous findings in women scheduledto undergo elective cesarean delivery during SA.9,10

Chamchad et al.9 demonstrated the predictive value ofHRV for hypotension accompanying SA for cesareandelivery. Point correlation dimension (PD2), a measureof HRV, was investigated, and significant differencesbetween hypotensive and normotensive patients weredemonstrated. Recently, we showed the predictive valueof HRV for hypotension after SA in obstetric patients. Itwas demonstrated that LF/HF 2.5 correlated signifi-cantly with hypotension. The same LF/HF level intro-duced in pregnant women was defined to discriminate

low and high risk of hypotension in elderly men aimingat a simple measure of risk evaluation applicable todifferent groups of patients.10 We conclude that LF/HFof 2.5 may be a cutoff value independent from underly-ing individual conditions. To evaluate its predictivevalue, preoperative LF/HF was correlated with the de-gree of SBP decrease after SA. A significant correlationwas demonstrated, and a high sensitivity and specificityof LF/HF 2.5 to predict hypotension was shown. Inaddition, baseline HF (reflecting vagal activity) was cor-related with the percentage decrease of SBP. A signifi-cant correlation was demonstrated confirming the viewthat human cardiac function is mainly controlled viavagal mechanisms.26 Nevertheless, receiver operatorcurve analyses of LF/HF demonstrated better sensitivityand specificity compared with receiver operator curveanalysis of HF. Better sensitivity and specificity of LF/HF

Table 3. HRV Data

Event HRV MILD MOD SEV LF/HF 2.5 LF/HF 2.5

DBSLF, % 30 (14/52) 53 (25/69) 36 (25/58) 35 14 34 12HF, % 49 (24/62) 27 (12/54) 26 (12/48) 35 19 26 17LF/HF 0.7 (0.3/2.0) 1.8 (0.7/3.1) 2.4 (0.7/4.4) 1.5 1.2 2.0 1.5ToPo, ms2 138 (80/369) 406 (169/496) 523 (255/891) 1,820 2,400 460 420Mean R-R, s 0.8 (0.65/1.3) 1.0 (0.7/1.4) 0.9 (0.7/1.1) 0.9 0.2 0.9 0.1SD R-R, ms 25 (15/36) 35 (15/36) 35 (25/74) 47 34 33 12

DOS-BLLF, % 23 (15/37) 35 (22/67) 39 (34/57)* 31 13 51 14HF, % 39 (15/50) 12 (9/21)* 5 (3/6)*/ 36 22 11 5LF/HF 0.7 (0.4/1.2) 3.0 (1.4/4.2)* 7.5 (4.1/11.0)*/ 1.2 0.8 5.5 2.5ToPo, ms2 253 (155/484) 379 (172/691) 393 (132/449) 1,600 1,090 480 430Mean R-R, s 0.8 (0.7/1.2) 0.8 (0.7/1.1) 0.8 (0.6/0.9) 0.9 0.2 0.8 0.1SD R-R, ms 33 (22/96) 33 (19/58) 32 (17/69) 107 58 34 14

DOS-PRELF, % 26 (16/30) 41 (22/67)* 43 (34/57)* 28 12 38 11HF, % 29 (23/36) 24 (10/32) 14 (7/17)* 39 23 16 11LF/HF 1.0 (0.5/1.2) 1.4 (0.7/5.2) 3.5 (1.4/6.9)* 1.3 1.0 3.9 3.0ToPo, ms2 176 (140/680) 134 (104/416) 325 (89/434) 1,550 1,070 490 375Mean R-R, s 0.9 (0.6/1.3) 0.8 (0.7/1.2) 0.8 (0.7/1.3) 0.9 0.1 0.8 0.1SD R-R, ms 33 (18/83) 26 (13/50) 26 (14/58) 60 56 32 17

DOS-SA15LF, % 34 (18/42) 29 (10/43) 28 (19/38) 23 14 31 12HF, % 38 (24/52) 14 (8/24) 21 (7/41) 49 22 24 17LF/HF 0.8 (0.5/1.7) 2.1 (0.6/5.0) 2.6 (0.7/5.7) 0.8 0.6 2.2 1.9ToPo, ms2 296 (165/2,149) 285 (168/534) 445 (221/657) 2,130 1,930 470 410Mean R-R, s 0.9 (0.7/1.2) 0.9 (0.8/1.3) 0.9 (0.7/1.1) 1.0 0.2 0.9 0.2SD R-R, ms 25 (11/115) 28 (14/43) 40 (27/61) 65 51 32 16

DOS-SA45LF, % 23 (14/27) 25 (15/32) 18 (10/29) 23 12 29 13HF, % 45 (22/58) 19 (10/27) 35 (8/41) 48 22 22 15LF/HF 0.6 (0.4/1.4) 1.3 (1.0/2.3) 0.7 (0.3/4.0)* 0.9 0.7 2.2 2.0ToPo, ms2 904 (351/1516) 620 (271/999) 532 (188/840) 2,200 2,100 850 690Mean R-R, s 1.0 (0.8/1.3) 1.0 (0.8/1.2) 0.9 (0.8/1.1) 1.0 0.2 1.0 1.0SD R-R, ms 52 (9/100) 32 (13/56) 52 (21/86) 69 52 45 19

Retrospective groups: Data are presented as median and minimum/maximum. Prospective groups: Data are presented as mean SD.

* P 0.05 vs.MILD. P 0.05 MOD vs. SEV. P 0.05 LF/HF 2.5 vs. LF/HF 2.5. P 0.05 within group in the course of spinal anesthesia comparedwith DOS-BL.

DBS day before surgery; DOS-BL day of surgery baseline; DOS-PRE day of surgery after prehydration; DOS-SA15 measurement 15 min after spinalanesthesia; DOS-SA45 measurement 45 min after spinal anesthesia; HF high frequency; HRV heart rate variability; LF low frequency; LF/HF lowto high frequency; LF/HF 2.5 prospective group with low baseline low to high frequency ratio; LF/HF 2.5 prospective group high baseline low to highfrequency ratio; Mean R-R mean of consecutive beat-to-beat intervals; MILD retrospective group with mild hypotension; MOD retrospective group withmoderate hypotension; SD R-R SD of mean R-R interval; SEV retrospective group with severe hypotension; ToPo total power.




Fig. 2. Heart rate variability, retrospective groups. (A) Low tohigh frequency ratio (LF/HF). (B) Low frequency (LF). (C) Highfrequency (HF). DBS day before surgery; DOS-BL day ofsurgery baseline; MILD mild hypotension; MOD moderatehypotension; PRE after prehydration; SA5 5 min afterspinal anesthesia; SA15 15 min after spinal anesthesia;SA45 45 min after spinal anesthesia; SEV severe hypoten-sion. Data are presented as median, 25th and 75th percentile,and range. * P < 0.05 versus MILD. # P < 0.05 versus SEVERE. P < 0.05, significant difference within groups between dif-ferent events.

Fig. 3. Heart rate variability, prospective groups. (A) Low to highfrequency ratio (LF/HF). (B) Low frequency (LF). (C) High fre-quency (HF). DBS day before surgery; DOS-BL day ofsurgery baseline; LF/HF < 2.5 baseline LF/HF less than 2.5;LF/HF > 2.5 baseline LF/HF greater than 2.5; PRE afterprehydration; SA5 5 min after spinal anesthesia; SA15 15 min after spinal anesthesia; SA45 45 min after spinalanesthesia. Data are presented as mean SD. * P < 0.05 versusLF/HF > 2.5. P < 0.05, significant difference within groupsbetween different events.

542 HANSS ET AL.



may indicate that sympathetic control reflected by LFaffects cardiac regulation as well. Therefore, the ratio oflow to high frequency, reflecting sympathetic to para-sympathetic balance, is superior to HF alone. Time do-main parameters mean R-R interval and SD of R-R interval

did not differ significantly between groups depending onpostspinal hypotension. These parameters mainly reflectparasympathetic regulation. Therefore, frequency do-main analysis, especially the LF/HF ratio, may be moresensitive for prediction of postspinal hypotension com-

Fig. 4. Correlation of baseline parameters and hypotension. (A) Correlation of low to high frequency (LF/HF) and decrease of systolicblood pressure (SBP). (B) Correlation of high frequency (HF) and decrease of SBP. (C) Correlation of heart rate (HR) and decreaseof SBP. (D) Correlation of baseline (BL) SBP and decrease of SBP. r > 0.5, P < 0.05.

Fig. 5. Receiver operator curve (ROC) analyses. (A) ROC of low to high frequency ratio (LF/HF). (B) ROC of high frequency (HF).Area area under the curve. P < 0.05.




pared with parameters solely reflecting parasympatheticcontrol.

There is evidence that baseline HR and SBP may indi-cate the risk of post-SA hypotension.19,2729 Therefore,these parameters were correlated with hypotension inour study. In contrast, we found no correlation betweenhemodynamic parameters and hypotension and onlyweak sensitivity and specificity for prediction of hypo-tension of HR and SBP in our study. Therefore, differentlevels of LF/HF, which may reflect sympathovagal bal-ance, correlate with comparable baseline hemodynamicparameters. HRV analysis may be superior comparedwith baseline HR or SBP for risk stratification of post-SAhypotension.

The sympathovagal balance is the result of sympatheticas well as parasympathetic influences. Basically, HRVanalysis is based on measuring the beat-to-beat interval ofthe sinus rhythm. These changes vary based on manyinfluences, such as volume status, respiration, intratho-racic pressure, and baroreceptor reflexes. If LF/HF mayreflect sympathovagal balance, it is an interesting ques-tion whether changes of LF, HF, or both result in mod-ifications of LF/HF. Only baseline HF, not LF, was signif-icantly different between retrospectively definedgroups. In prospectively assigned patients, both param-eters, LF as well as HF, were significantly different atbaseline, resulting in a significant difference of LF/HF.HF and vagal activity seem to be more clearly associatedthan LF and sympathetic activity.2023,30,31 Burgos etal.30 investigated ANS regulation in patients with under-lying autonomic reflex dysfunction due to diabetes mel-litus. Various tests, including HRV, were performed toinvestigate autonomic nervous system regulation. It wasdemonstrated that patients with impaired ANS regula-tion were at high risk of hypotension after induction ofgeneral anesthesia. Especially impaired parasympatheticactivity, reflected by HF, indicated a high risk of hemo-dynamic instability. Latson et al.31 investigated auto-nomic reflex dysfunction in patients presenting for elec-tive surgery for its prediction of hypotension. Theauthors suggested that indicators of vagal activity, in-cluding HF of HRV, were best correlated with hypoten-sion after induction of general anesthesia. Our studyinvestigated LF/HF as the resulting ratio of both param-eters, LF as well as HF, and HF alone. LF/HF demon-strated better sensitivity and specificity for prediction ofhypotension compared with isolated HF, indicating thatsympathetic activity contributes to cardiac regulation aswell.

HRV in the Course of SAHypotension during SA is mainly a result of decreased

systemic vascular resistance caused by inhibition ofpreganglionic sympathetic nerve fibers. High thoracicblockade may in addition block sympathetic innervationof the heart by blocking nervi accelerantes, resulting in

bradycardia and decrease of stroke volume. The coinci-dence of a decrease of LF with hypotension was demon-strated in patients during SA. SA provoked a decrease ofLF and a small increase of HF resulting in a decrease ofLF/HF.32 These changes of HRV during the course of SAwere comparable to our results. LF/HF and LF decreasedafter SA accompanied by hypotension. In contrast, pa-tients who showed low baseline LF/HF and LF as well asincreased HF did not have development of significantchanges of HRV after SA, reflected by only minorchanges of hemodynamic parameters. The importanceof the balance of LF/HF was previously demonstrated.33

A depression of LF as well as HF in patients undergoingcesarean delivery during SA or epidural anesthesia wasdemonstrated. Consequently, LF/HF remained un-changed, accompanied by stable hemodynamic parame-ters. We conclude that blockade of preganglionic sym-pathetic fibers and blockade of sympathetic innervationsof the heart are associated with a decrease of LF/HF andLF. These findings suggest that LF/HF and LF may beindirect measures of the sympathetic activity of the ANSregulation.

LimitationsSome limitations of our study should be noted. HRV is

only an indirect measure of autonomic activity. It wasestablished in this clinical setting because noninvasivedirect measures of autonomic regulation are not avail-able. Artifacts during HRV data recording were inevita-ble to some degree. However, artifacts were eliminatedby computer-based artifact detection followed by anevaluation by an expert blinded to the hemodynamiceffects of SA. Because fast Fourier transformation analy-sis requires stationary data, patients were asked to liecalmly in the supine position during measurements. Be-cause of the minimally invasive technique of the surgicalprocedure with almost no blood loss and minimal surgi-cal manipulations, stillness was secured throughout themeasurements. Vasopressor boluses after SA for treat-ment of hypotensive episodes may influence HRV mea-surements in the course of SA.

Heart rate variability measurements are based on stan-dard electrocardiographic recordings at high acquisitionrates, and commercial tools offer computerized interpre-tation. Therefore, in principle, the technique could eas-ily be implemented in routine clinical monitoring. HRVmeasurements may contain important prognostic infor-mation about individual hemodynamic reaction after SA.Our data suggest a correlation between an increasedsympathetic activity indicated by HRV before SA andhypotension in the course of regional anesthesia. A highsensitivity and specificity of LF/HF greater than 2.5 forprediction of hypotension was demonstrated. Therefore,HRV analysis before SA may be suitable to detect patientsat high risk of hypotension. The predictive value ofLF/HF may be superior to established predictors.

544 HANSS ET AL.



References

1. McKenzie PJ, Wishart HY, Dewar KM, Gray I, Smith G: Comparison of theeffects of spinal anaesthesia and general anaesthesia on postoperative oxygen-ation and perioperative mortality. Br J Anaesth 1980; 52:4954

2. Brown AG, Visram AR, Jones RD, Irwin MG, Bacon-Shone J: Preoperativeand postoperative oxygen saturation in the elderly following spinal or generalanaesthesia: An audit of current practice. Anaesth Intensive Care 1994; 22:1504

3. Olofsson C, Nygards EB, Bjersten AB, Hessling A: Low-dose bupivacainewith sufentanil prevents hypotension after spinal anesthesia for hip repair inelderly patients. Acta Anaesthesiol Scand 2004; 48:12404

4. Hartmann B, Junger A, Klasen J, Benson M, Jost A, Banzhaf A, HempelmannG: The incidence and risk factors for hypotension after spinal anesthesia induc-tion: an analysis with automated data collection. Anesth Analg 2002; 94:15219

5. Brooker RF, Butterworth JFT, Kitzman DW, Berman JM, Kashtan HI, McK-inley AC: Treatment of hypotension after hyperbaric tetracaine spinal anesthesia:A randomized, double-blind, cross-over comparison of phenylephrine and epi-nephrine. ANESTHESIOLOGY 1997; 86:797805

6. Lee A, Ngan Kee WD, Gin T: A quantitative, systematic review of random-ized controlled trials of ephedrine versus phenylephrine for the management ofhypotension during spinal anesthesia for cesarean delivery. Anesth Analg 2002;94:9206

7. Bootsma M, Swenne CA, Van Bolhuis HH, Chang PC, Cats VM, Bruschke AV:Heart rate and heart rate variability as indexes of sympathovagal balance.Am J Physiol 1994; 266:H156571

8. Pomeranz B, Macaulay RJ, Caudill MA, Kutz I, Adam D, Gordon D, KilbornKM, Barger AC, Shannon DC, Cohen RJ: Assessment of autonomic function inhumans by heart rate spectral analysis. Am J Physiol 1985; 248:H1513

9. Chamchad D, Arkoosh VA, Horrow JC, Buxbaum JL, Izrailtyan I, Nakham-chik L, Hoyer D, Kresh JY: Using heart rate variability to stratify risk of obstetricpatients undergoing spinal anesthesia. Anesth Analg 2004; 99:181821

10. Hanss R, Bein B, Ledowski T, Lehmkuhl M, Ohnesorge H, Scherkl W,Steinfath M, Scholz J, Tonner PH: Heart rate variability predicts severe hypoten-sion after spinal anesthesia for elective cesarean delivery. ANESTHESIOLOGY 2005;102:108693

11. Eneroth E, Westgren M, Ericsson M, Lindblad LE, Storck N: 24-hour ECGfrequency-domain measures in preeclamptic and healthy pregnant women dur-ing and after pregnancy. Hypertens Pregnancy 1999; 18:19

12. Lewinsky RM, Riskin-Mashiah S: Autonomic imbalance in preeclampsia:Evidence for increased sympathetic tone in response to the supine-pressor test.Obstet Gynecol 1998; 91:9359

13. Task Force of the European Society of Cardiology and the North AmericanSociety of Pacing and Electrophysiology: Heart rate variability: Standards ofmeasurement, physiological interpretation and clinical use. Circulation 1996;93:104365

14. Polanczyk CA, Rohde LE, Moraes RS, Ferlin EL, Leite C, Ribeiro JP: Sym-pathetic nervous system representation in time and frequency domain indices ofheart rate variability. Eur J Appl Physiol Occup Physiol 1998; 79:6973

15. Uusitalo AL, Tahvanainen KU, Uusitalo AJ, Rusko HK: Non-invasive evalu-ation of sympathovagal balance in athletes by time and frequency domain anal-yses of heart rate and blood pressure variability. Clin Physiol 1996; 16:57588

16. Hirsch JA, Bishop B: Respiratory sinus arrhythmia in humans: How breath-ing pattern modulates heart rate. Am J Physiol 1981; 241:H6209

17. Liu SS, McDonald SB: Current issues in spinal anesthesia. ANESTHESIOLOGY2001; 94:888906

18. Hanley JA, McNeil BJ: The meaning and use of the area under a receiveroperating characteristic (ROC) curve. Radiology 1982; 143:2936

19. Frolich MA, Caton D: Baseline heart rate may predict hypotension afterspinal anesthesia in prehydrated obstetrical patients. Can J Anaesth 2002; 49:1859

20. Akselrod S, Gordon D, Ubel FA, Shannon DC, Berger AC, Cohen RJ: Powerspectrum analysis of heart rate fluctuation: A quantitative probe of beat-to-beatcardiovascular control. Science 1981; 213:2202

21. Pagani M, Lombardi F, Guzzetti S, Rimoldi O, Furlan R, Pizzinelli P,Sandrone G, Malfatto G, DellOrto S, Piccaluga E: Power spectral analysis of heartrate and arterial pressure variabilities as a marker of sympatho-vagal interaction inman and conscious dog. Circ Res 1986; 59:17893

22. Malliani A, Pagani M: Spectral analysis of cardiovascular variabilities in theassessment of sympathetic cardiac regulation in heart failure. Pharmacol Res1991; 24 (suppl 1):4353

23. Montano N, Ruscone TG, Porta A, Lombardi F, Pagani M, Malliani A: Powerspectrum analysis of heart rate variability to assess the changes in sympathovagalbalance during graded orthostatic tilt. Circulation 1994; 90:182631

24. Hopf HB, Skyschally A, Heusch G, Peters J: Low-frequency spectral powerof heart rate variability is not a specific marker of cardiac sympathetic modula-tion. ANESTHESIOLOGY 1995; 82:60919

25. Kawase M, Komatsu T, Nishiwaki K, Kobayashi M, Kimura T, Shimada Y:Heart rate variability and arterial blood pressure variability show different char-acteristic changes during hemorrhage in isoflurane-anesthetized, mechanicallyventilated dogs. Anesth Analg 2002; 94:1621

26. Azevedo ER, Parker JD: Parasympathetic control of cardiac sympatheticactivity: Normal ventricular function versus congestive heart failure. Circulation1999; 100:2749

27. Kinsella SM, Norris MC: Advance prediction of hypotension at cesareandelivery under spinal anesthesia. Int J Obstet Anesth 1996; 5:37

28. Berlac PA, Rasmussen YH: Per-operative cerebral near-infrared spectros-copy (NIRS) predicts maternal hypotension during elective caesarean delivery inspinal anaesthesia. Int J Obstet Anesth 2005; 14:2631

29. Ouzounian JG, Masaki DI, Abboud TK, Greenspoon JS: Systemic vascularresistance index determined by thoracic electrical bioimpedance predicts therisk for maternal hypotension during regional anesthesia for cesarean delivery.Am J Obstet Gynecol 1996; 174:101925

30. Burgos LG, Ebert TJ, Asiddao C, Turner LA, Pattison CZ, Wang-Cheng R,Kampine JP: Increased intraoperative cardiovascular morbidity in diabetics withautonomic neuropathy. ANESTHESIOLOGY 1989; 70:5917

31. Latson TW, Ashmore TH, Reinhart DJ, Klein KW, Giesecke AH: Autonomicreflex dysfunction in patients presenting for elective surgery is associated withhypotension after anesthesia induction. ANESTHESIOLOGY 1994; 80:32637

32. Kimura T, Komatsu T, Hirabayashi A, Sakuma I, Shimada Y: Autonomicimbalance of the heart during total spinal anesthesia evaluated by spectralanalysis of heart rate variability. ANESTHESIOLOGY 1994; 80:6948

33. Landry DP, Bennett FM, Oriol NE: Analysis of heart rate dynamics as ameasure of autonomic tone in obstetrical patients undergoing epidural or spinalanesthesia. Reg Anesth 1994; 19:18995




HRV to Predicts Hypotension After Spinal Anesthesia

Documents

Transcript of HRV to Predicts Hypotension After Spinal Anesthesia