Effects of three anesthetic induction techniques on heart rate variability

Original Contributions Effects of Three Anesthetic Induction Techniques on Heart Rate Variability Terry W. Latson, MD,* S. Maire McCarroll, MB, BCh, FFARCSI,? M. Andrew Mirhej,$ Vernon A. Hyndman,§ Charles W. Whitten, MD,11 James M. Lipton, MD# Department at Dallas,

*Associate

Professor

tvisiting Assistant Anesthesiology SMedical

Student

OBiomedical (IAssistant

of Anesthesiology Professor

of

III

Engineering Professor

Specialist

of Anesthesiology

#Professor of Physiology, Research ciate Professor of Anesthesiology

Asso-

Address reprint requests to Dr. Latson at the Department of Anesthesiology, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas,

TX 75235-8894, USA. Dr. Latson was supported in part by an Institutional Research Grant from Southwestern Medical School. Parts of this material were presented at the Annual Meeting of the American Society of Anesthesiologists, Las Vegas, October 1% 23, 1990. Received for publication September 5, 199 1; revised manuscript accepted for publication January 10, 1992. 0 1992 Butter-worth-Heinemann J. Clin. Anesth.

4:265-276,

1992.

of Anesthesiology, Dallas,

Cniversity

of Texas

Southwestern

Medical

Center

TX.

Study Objective: To investigate the effects of diff erent clinical induction techniques on heart rate variability (HRV). Design: Two studies are reported. Study 1 prospectively compared the effects of two induction techniques (etomidate vs. thiopental sodium) known to have widely disparate effects on cardiovascular reflexes. Study 2 specifically investigated whether the vagotonic effects of sufentanil cause an increase in vagally mediated HRI’. Setting: Electi7le .surgery in a university-affiliated hospital. Patients: Study I: 18 ASA physical .ctatus I patients having minor surgery; Study 2: 10 ASA physical status III and IV patients having cardiac surgery. Interventions: In Study I, anesthesia was induced with either etomidate 0.3 mglkg or thiopental sodium 4 mglkg with 6OYo nitrous oxide in oxygen. In Study 2, anesthesia was induced with a sufentanil infusion (total dose 2.9 * 0.2 pglkg). Measurements and Main Results: The electrocardiogram-derived heart rate signal was subjected to power spectral analysis (similar to electroencephalopraphic analysis) to obtain measurements of (1) absolute HRV power [units of (beats per minute}21 within defined frequency ranges (HRV co = power between 0 and 0.125 Hz; HRV,, = power between 0.126 and 0.5 Hz; HRV,, = HRV,, f HRV,,) and (2) normalized HRV power (the percentage of total power) within these same frequency ranges [e.g., %HRVn, = (HRVJHRV,,) x 1 OO%]. In Study 1, both techniques caused large reductions in HRV,,. The reduction caused by the thiopental sodium technique ( - 89% 2 2%) significantly exceeded that caused by the etomidate technique ( - 58% 5 13%, p < 0.02). In Study 2, sufentanil decreased absolute power measurements of vagally mediated HRV (- 69 2 12 change in HRV,,) but increased corresponding normalized measurements of vagally mediated HRV (90%> 2 30% increase in %HR V,,). Conclusions: In Study I, the greater reduction in HRV with the thiopental sodium technique provides evidence that the depressant effects of anesthetics on HRV are related in part to their effects on cardiovascular reflexes. However, th.e significant depression in HRV caused by the etomidate technique suggests that mechanisms other than baroreflex depression (e.g., impaired consciousness) also are important in these depressant effects. In Study 2, the decrease in HRV,, caused by sufentanil documents that absolute power measurements of vagally mediated HRV are not correlated with changes in J. Clin. Anesth.,

vol. 4, July/August

1992

265

Original Conttibutions

Keywords: Anesthesia,

general; anesthetics; autonomic nervous system; baroreflexes; etomidate; heart rate: sufentanil; thiopental sodium.

Introduction I’kie term

heart r-de ztaridnlity (HKV) describes small oscillations in heart rate (HK) caused by the interaction of’ multiple regulatory influences on the sinus node. Power spectral analysis of’ HKV can provide noninvasive measurements of‘ autonomic nervous system activity.’ i ‘I‘his analysis measures both the amplitude and frequency of’ these HK oscillations [similar to power spectral analysis of’ the electroencephalogram (EEG)]. Oscillations occurring at dif‘ferent frequencies are associated with the acLivity of‘ different autonomic reflexes. Oscillations occurring at frequencies higher than about 0.123 Hz (i.(,.. a repetirion period OF 8 seconds or less) are mediated primarily by parasympathetic reflexes associated with sinus arrhythmia).’ L respiration (t.~., the respiratory Changes in the amplitude of’ these high-frequency oscillations have been correlated with changes in parasyrnOscillations occurring at lowetpathetic activity.& l’requencies at-e thought to be tnediared predominarttl! reflexes,‘,’ although pharmacologic by sympathetic blocking studies suggest that parasympathetic reflexes also are involved. Changes in the amplitude of’these lowf’requency oscillations have been cot-related with changes in sympathelic activit)..‘.’ HKV analysis may have itnportanr applications to the study of‘al~erations in autonomic activity caused by anes1hesia and surgery. Prior studies in the anesthesia literature have related alterations in HKV to both depth of‘ anesthesia”-” and changes in central autonomic outflow.*lC’ 1’However, important questions retnain regarding the application of’HRV analysis to anesthesia studies. One essential question is whether dif’f’erent anesthetic drugs have different effects on HKV. All prior studies isoflurane,“’ enHurane,* with anesthetics (halothane,” thiopental sodium, 11and the combination sevoflurane,li of’ fentanyl midazolam, and pancuronium*) have shown reductions in all f‘requency components of- HKV that wet-e measured during general anesthesia. However, all of these studies were limited to single anesthetic techniques, they were performed in varied study populations, and they used different analytical methods. Making meaningf’ul contparisons of’ the effects of‘ different an-

266

J. Clin. Anesth.,

vol. 4, July/August

1992

es1 hrtic-s on the basis 01’these pre\ ious srudies is. therr>fore, dif‘ficult. :\tiothet~ itnportattf and related queslion i5 the eliofog!’ of the reductions in HKV caused bv anesthetic-~ Prior studies itt t~rtNttl~.sthe/izud human sttbjects have related changes in HKV to several mechanisms: alteratiotr~ in svmpathetic and parasympathetic activit!., ’ ’ changes itt baroreflex sertsitivit\,.” I1 tmpaired cerebral function (resulting f’rom brain injury), Ii I0 interrupted au1ononiic neuropath~~.‘” 1” TOW reflex pathwavs. I7 and autonomic. extent to which each of‘ these mechanisms tnav be itt\,olvetl with anesthetic effects art HKV is ttnc-lea;.. If’ im l~unc tion associated witlt the loss 01 paired cerebral ~~otis~~iottsttess c-;ittsed b\ ~general attesthrsia ltas
Materials

cottipottetits

01‘ FIR\:.’

and Methods

i\l.trr approval b) llte l:ttiL,et-sity 0t“l’cxas Southwt3Lerrt Lledical Center‘s Institutional Keview Board. WC e\,alttaled the ef‘feccs of’ comparable thiopental sodium mtl etomidatt. inductions on HR\’ in IH ,4SA physical status I patients scheduled for minor elective surgery. Inti)rmetf consen was obtained front all patients. Subjecrs ranged in age from 18 to 35 years (mean 24 years) and itt weighl from 46 to W kg. Patients were assigned to receive either thiopental sodium or erotnidate on an alternating basis. Prentedication consisted of an intrarettous (LV) dose of’sut~ntanil 0. I5 kg/kg given upon arrival in I he operating room (OK). ‘I-hit-t\, seconds priot. to induction, all patients recei\:ed 50 mg of’ IV lidocaine. This drug was given to prevent painful burning in the patients who were to receive etomidate: it was given to

Anpsthetic effects on heart rate variability:

all patients to ensure comparable study conditions. Patients were then given 4 mg/kg of thiopental sodium (n = 9) or 0.3 mgikg of etomidate (n = 9). After loss of consciousness, patients were paralyzed (vecuronium 0.1 ma/kg or atracurium 0.5 mg/kg) and their lungs ventilated by mask with 60% N,O in oxygen (0,). Except for mask ventilation, patients were left unstimulated for a period of 3 to 4 minutes after induction. Postinduction measurements were collected during this interval. Changes in respiratory pattern can cause changes in high-frequency HRV. Similar to other studies, no attempt was made to control ventilation during baseline measurements of HRV. Following induction, all subjects were ventilated at a rate of approximately 12 breaths per minute with normal tidal volumes (assessed by chest wall excursion). This respiratory frequency (0.2 Hz) was selected to keep the resultant respiratory variations in HR within the defined frequency range for our measurements of high-frequency HRV. Electrocardiogram (EC(;) recordings for HRV analysis were obtained using a computer equipped with a 12bit resolution analog-to-digital converter. The ECG signal from the OR monitor was sampled at the analog output connector at a rate of 250 Hz. Preinduction recordings wei-e begun as soon as possible after the patient entered the OR. Postinduction recordings commenced with injection of thiopental sodium or etomidate and continued for at least 3 minutes. ‘l‘he steps involved in deriving the HRV power spectral measurements are shown in F@LTP I. The first step was to locate each QRS complex in the recorded EGG signal. .I‘bis was done using a cross-correlation technique, in which a QRS template was aligned with each detected QRS complex. .l‘his method ensures that a constant reference point within each complex is used for determination of the R-to-R intervals. A “continuous” HR signal (sampled at 4 Hz) was then constructed from these R-to-R intervals using the methods described by Berger rt CC~.“ Splining ~ techniques (Le., linear interpolation be-

Aulo~orrelat~on

Function

(ACF)

Ei

I

I

Flltered 64 second remove

denvc

Figure

,mear

bend

Caussla”

AutocorrelatIon

1. Analytical

variability

HRV power

Spectrum

segments

super

apply

/

HR Signal

Fi w,rKkov

Function

methods f’or derivation (HKV) power measurements.

of heart rate

Latsorz et al.

tween valid data points) were used to remove any artifactual effects of ectopic beats on this HR signal.” Since the effects of bolus injections of the study drugs were relatively short-lived, the length of the data segments used for HRV analysis was necessarily relatively short. A segment length of 64 seconds was selected. This short segment length precludes accurate assessment of very low frequency oscillations (period lengths more than 32 seconds). Any such oscillations may introduce artifacts in the subsequent analysis. To avoid these potential artifacts, the HR signal was prefiltered to remove signal components at frequencies lower than 0.0312 Hz using standard digital filtering techniques (band pass filter, 0.0312 to 0.5 Hz). Spectral analysis was then performed on the filtered HR signal. The steps involved in this analysis were as follows: (1) Any residual linear trend in the 64-second data segment was removed, and the ends of the data segment were tapered using a Super Gaussian window. These steps were taken to reduce the artifacts introduced in the fast Fourier transform (FFT) when the starting and final values of a signal are not equal (termed leak~ge).i’~ (2) The autocorrelation function for this tapered data segment was calculated using the FFT. (3) The ends of the autocorrelation function were tapered using another Super Gaussian window. This step has the effect of smoothing the resulting HRV power spectrum (similar IO averaging).‘” (4) The HRV power spectrum, which is the Fourier transform of the autocorrelation function, was then calculated using the FFT. The HRV power spectrum [units of (beats per minute, bpm)?/Hz] describes HRV power as a function of frequency (Hz). Based on conventions in the literature, HRV parameters are typically reported as HRV power contained within specific frequency, ranges. Measurements of’ ctbsolutr HRV power (units of bpm’) are obtained by integration of the area under the power spectral curve over the defined frequency range. Absolute power measurements represent the variance (standard deviation squared) of the HR about a given mean. HRV parameters also are often reported in terms of rlormnlized power measurements. ‘l‘hese measurements represent HRV power in a specific frequency range expressed as a proportion of total HRV power (f.~., area under the curve for a specific frequency range divided by area under the curve for all frequencies). The most commonly reported HRV parameters are low-frpqurncy pouler and high-frequency ,L~Lw. Although the precise frequency ranges associated with these parameters have varied slightly among authors, typical values are 0.03 to 0.125 Hz for the former and 0.126 to 0.5 Hz for the latter. ‘Phese ranges are based on the locations of characteristic peaks in the frequency spectrum associated with the activity of different reflexes. In the present study, specific HRV measurements calculated for each 64-second epoch included the following: (1) low-frequency power (HRV,.,,; the area under the power spectral curve between 0.03 and 0.125 Hz); (2) high-frequency power (HRV,,,; the spectral area between 0.126 Hz and 0.5 Hz); (3) total HRV power the sum of HRV,,, and HRV,,,); (4) normal(HRV,,,,; J. Clin. Anesth..

vol. 4, July/August

1992

267

ized low-frequency power [SHKV,,, = (HKV,,,/HKV,,,,) X lOO%]: (5) normalized high-fi-equency power [S’HKV,,, = (HKV,,,lHKV,,,,) x IOWl. ‘I*rended measurements of’ HKV ~‘et-e derived fq SCqwntial analysis of overlapping (i4-second data segmenls, with each sequential 64-second segment heginning of’ the previous segment. 15 seconds af’ter the beginning Example results are shown in F~~qw~~2. The sequential HKV spectra are plotted as a tllree-tfinlensic,rl;ll surfilcc that describes changes in HKV power over lime (similar to cwmpressed spectral array plots of‘thv WC). ln~egralion of sped i-al values over the appropriate frequencies wig3 provides corresponding 1rentted tiieasuremen~s 01’ rhe f‘i.ccficeiic.?,-spec.iflc, p;lr;mielei.s lisletl ahove. Exatnple trends of‘ HKV, () mtl HKV,,, ;tre show11 in FjCcm (0.P 3. Pi-&itluctioii values 01’ HKV wtw ohcaiiietl by aver-aging the f‘ive sequential trended HKV measurenwn~s obtained just prior to induction (tort-esI”“i(lingI0 ii 124sfxoiitl time period). ‘I‘he maximal ef’l‘ec-t of’the induction drugs on ~rentled HKV measLlremen~s t~suallv occurred lwt~wri 90 md 180 seconds af’ter dl-rig ;itln;inisll-atit,II (I;iq”, 3) mtl w;is 01’variable duralion. ‘l’hc poatintltrc~lion nirasut-rnients used for comparisotl with pwiiiduv tioil ineasuwnients wert‘ oblaiiieti hv ;i\vraging rliwc 1rrndetl ilieasLii.t~iii~~litb tlu&ig the period of seqtienfial m~xintal reduction in HKV in each pa’ient. 111 (his stutl~. u’c irl\cstigated changes in HKC’ during induction of’;~ncsttiesi;i with strfentanil inf’usion. WV cxatnineti both steady-stare ch;mges in HKV l>elwt‘t‘Il tlw preincfuc tion and poslinditc.tion periods and lrcndcd changes in HKV during rhc gdual loss oJ‘c.onscio~rsrIess c~iiisetl by suf~ntanil. ‘I‘his study wxs appiu~ed by the Ilniversilv of‘ .l‘exas Southwestern Medic-al (;enter’r Instittltionitl Keview Roar-d, and infornd rollsrnt was ohl‘he study group consisted 01 r;linetl f’rom all p’ients. Ien :\S.\ physical status If1 and IL’ patients scheduled fi)r electiw cardiac surger! (seven fitr coromry arter\

lymss.

Iwo hi- mitt-al valve replaceriieir~,

surgical ;iblation of’ ;iIl (Ltrcliac surgid parients (.;uise high-dose sufbntanil

mtl

oiie

foi

wnducrion p;iIhw;ty), were selecreti for this sttitl~~ tw tisrci inductions dre cwrirnonl~

acu3sor~~

fol. Ihew p’irnts. Shjeccs ranged iii age I’i-om 42 10 70 years (mwn 54 vmrs) ‘and in weight from 51 to 100 kg.

M’irh resped to preopelariw c~arcliac inetlic;~lions, eight patic:nts were taking c-al&ml channel blockers. thrw wrt raking ;ulgiotensin converting enzwne inhibitors. one uas

Figure I wrItled

hearr rate (HK) signal aid 2. High-resolution heart rate variability (HKV) power spec’t ra. l’owct.

spect I.21 :11-e spaced 15 seconck apart. EXll specr 1‘11111 w:Is derived f’rom ;L G&second data epoch of the HK signal (:I2 srcontls on 13~1~ sitle of’ the plotred spectr-a). ‘IXctperital sodium was administered at the time intiic~atetl (‘IT. (BPM F heats I>el. minute.) 268

J. Clin. Anesth.,

WI. 4, July/August

19Y2

taking irIsdin. and one \V;IStaking a beta Mocker. I’;uic.rlts wirh rriiclazolatn iliti-.;iIlilisC111;IL.1~ wwr prtwidicatetl (mt‘a~i dose 3.5 t 0.5 q) appwxiniately 43 miirules prior to al-A\;11 in the OK; atfditicmal doses of rnitlazolm~~ IV (~ireat~ 2.8 ir 0.6 nig) were adininisteretl 25 nectletl ti)r wcfation during ptatemenl of‘ radial arterial and pcriphrd I c’ (‘;*Iheter-s.

Anesthetic

hypnotic drugs were administered during induction. Level of consciousness was assessed every 15 seconds. For purposes of this study, loss of consciousness was defined as loss of responsiveness to loud verbal commands and no reaction to subsequent positive-pressure mask ventilation. After loss of consciousness, patients were paralyzed with vecuronium 10 mg IV, and their lungs were ventilated by mask with 100% 0, at a rate of 12 breaths per minute and normal tidal volumes (assessed by chest wall excursions). Patients were unstimulated until laryngoscopy. ECC recordings for HRV analysis and subsequent derivation of trended HRV measurements were performed in the same manner as in Study 1. Preinduction values of HRV were calculated by averaging five sequential trended measurements obtained just prior to induction. Postinduction values of HRV were obtained by averaging five trended measurements taken just prior to laryngoscopy. The significance of changes in HRV was assessed by the Wilcoxon signed rank test, with a value of /I s 0.05 considered significant.

Figure

effects on heart rate uariahility:

Letson

et al.

4.

High-resolution heart t-ate (HR) signals f*orn two Patient A received thiopental sodium and Note the reduction in amB received etomidatr. of’ HK oscillations afier drllg atlrllinistratiotl in both

study patients. Patient plitude patients.

(HPM = heats per minute.)

Results Study

change in HRV caused by thiopental sodium in this patient was similar to that observed in all other patients. Both induction techniques caused significant reductions in all measurements of absolute HKV power. -l’he effects of‘the two techniques are contrasted in Figurr 5. Compared with the etomidate induction, thiopental sodium induction caused a significantly greater reduction t 2% I,.\. - 58%’ f 13%, 1_ < in both HRV ,(), (-89% 0.02) and HRV,,, (-95% ? I’% I’.\. -47%;. X!I 13%. /I < 0.002). Although there was a trend toward thiopental sodium’s also having a greater depressant effect on HRV, 0, this trend did not reach statistical significance. Comparing the effects of each induction technique on the separate frequency components of‘ HRV, thiopental sodium caused a significantly greater reduction in HKV,,, than in HRV,,, (-95%# 2 1% I!.). -82%’ i: 4%,p < 0.01). In contrast, the effects of etomidate on these two measurements were not significantly different ( - 47% + 13%’ 7~. - 60% + 16%). As shown in Fipr-o 6, this differential effect of thiopental sodium on low- and high-frequency components caused significant changes in the normalized measurements of HRV (HRV,,,,: from 46% -t 6%

I

Representative HR tracings from patients in both induction groups are shown in Figure 4. Low-frequency and high-frequency HR oscillations are readily visible in these tracings. Patient A (top panel) received thiopenlal sodium, and patient B (botton panel) received etomidate. The timing of drug administration is noted. The reduction in amplitude of HR oscillations after drug administration is clear in both tracings. ‘I-he effects of these induction techniques on HRV were quantified using measurements of HRV power. Group mean values of HRV-,.,,-,., HRV,,,, and HRV,,, are shown in Tdd~ 1. One patient in the thiopental sodium group had an unusually high preinduction value of HRV,,, (38.74 bpm’), causing the mean preinduction value of HRV,,, in this group to be about twice the mean value in the etomidate group. However, whether this patient was included or excluded from the data analysis, there were no statistically significant differences in the preinduction values of HRV parameters between groups. Evaluated on a percent change basis, the subsequent

Table

1.

Heart Rate Variability (HRV) Power Measurements Thiopental

HKV r0, HKV,,,

HKV,,, “p -C 0.01 (preinduction **/I < 9.05 (preinduction

Sodium

helore

and after- Inducriou

(n = 9)

(n = 9)

Postinduction

Preinduction

Postinduction

22.33 10.18 12.15

1.85

14.11 x.41

6.44 3.71 2.72

k 4.93 k 2.39 k 3.54

k 0.411*

1.26 k 0.33* O.~iO -t 0. 19*

k 3.18 ” 2.43

5.69 k 1. 13

c 2.3!)” _c I .57** + 1.20””

US. postinduction). UT.postinduction).

Note: Units are (beats per minute)‘. Values are means t SEM. Statistics by HRV,,,,

Etomidate

Preinduction

= total HRV power; HRV ,() = low-frequency

Wilcoxon

signed

rank test.

HRV power; HRV,,, = high-trequenc)’

HRV ~mwer.

J. Clin. Anesth.,

vol. 4, July/August

1992

269

Or&id

Contributions

m

p< 02

p=NS -_

p” 002

H%l

HRV,

HR”,,

ThlopentaI

m

Etomldate

Figure 5. Effects of‘thiopental sodium and etomidate induction on absolute power measurements of’ heart rate WI.iability (HRV). Bars indicate the reduction in spectral power between preinduction and postinduction measurements. Brackets indicate SEM; significance level by Mann-Whitne) 17 test.

Figure 6. Effects oi thiopental sodium anti rtorniclatr on normalized power spectral measurements of‘heart rate \.a.iability (HRV). Bars indicate preinduction i’e/\/l.\ posrirlduction u~~~suremen~s. !&rackets indicate Sl%. :Isterisk denotes/)
to 71% _t 3%; HKV,,,: from 54%, -c- li(/; to 29%! t 4%). ‘I‘hiopental sodium caused a reduction in %‘HKV,,, in every patient. It caused an average 46% t 5% decrease in %HKV,,,, expressed as a percent change for each patient. Etomidate had no significant effect on normalized HKV measurements.

Shdy

2

An example

of- the effects of sufentanil on the derived HK signal and corresponding HRV spectral surface is shown in Figure 7. Sufentanil caused a marked reduction of’HKV power at all frequencies. The group mean values for preinduction and postinduction measurements of HKV are listed in Table 2. Suf-entanil caused an 83% 2 6%’ reduction in HKV,,,, (/J < O.Ol), an 88% t 5%’ reduction in HKV,,, (p < O.Ol), and a 69% 2 12%) reduction in HKV,,, (p < O.Ol), expressed as the average percent change from preinduction values. The reduction in HKV,,, was significantly less than the reduction in HKV,,,, (p < 0.01). As shown in Figure 8, this differential effect on low- and high-frequency components caused a significant change in normalized measurements of HKV. Sufentanil caused an increase in %HKV,, in nine out of ten patients, with the group mean value of %HKV,,, increasing almost twofold (from 24% ? 4%’ to 41% -+ 670, p < 0.01). Examination of the trended HKV measurements in individual patients revealed a consistent pattern: with loss of’consciousness, HKV decreased dramatically (see Figure 7). This sudden change was evident in all patients despite a wide variation in elapsed time (60 to 225 sec270

J. Clin. Anesth.,

vol. 4, July/August

1992

Figure 7. High-resolution heart rate (HR) signal and trended heart rate variability (HRV) power spectra during sufentanil induction. SIT denotes start of‘ suf’entanil infision. I,O(: denores loss of consciousness. HRV power at ~111 frequencies decreased following loss o~c.onsciousness, (BPM = beats per minute.)

ends) and sufentanil dose (1 to 2.5 Fgikg) preceding loss ofconsciousness. To quantify these changes in HKV with loss ofconsciousness, the trended measurements of HRV from the 64-second data segment ending just prior to loss of consciousness (pre-LOC) were compared with measurements from the 64-second data segment beginning just after loss of consciousness (post-1,OC). The group mean values f&- these measurements, along with the preinduction and postinduction measurements, are plotted in F&UP L).The large reductions in hoth HRV,,,

Anesthetic

effects an heart rate variabilit~~: Latson et al.

Table 2.

Heart Rate Variability (HRV) Power Measurements before and after Sufentanil Induction Pteinduction HRV,,,, I-IRV,.,, 11KV,,, *p < 0.0 1 (preinduction

:1.09 5 1.03 2.56 5 0.86 0.54 + 0.17

Postinduction 0.38 t 0.14* 0.26 t 0.10* 0.12 + 0.05*

ZLLpostinduction).

Note: Units are (beats per minute)‘. Values are means Statistics by Wilcoxon signed rank test.

2 SEM.

HRV,,,, = total HRV power; HRV,,, = low-frequency HRV power; HR\‘,,, = high-frequency HRV power.

Z

and HRV,, accompanying loss of consciousness are evident. As compared to the pre-LOC measurements, postLOC HRVTU, was reduced by 81% it_ 4% (p < O.Ol), post-LOC HRV,, was reduced by 81% -C 4% (fi < O.Ol), and post-LOC HRV,, was reduced by 65%’ 2 8% ($I < 0.01).

[-j

pre

% HRV,,

HRVLo

m

Induction

post InductIon

Figure 8. Kffects of. sufentanil on normalized power spectral measurements of heart rate variability (HRV). Bars indicate preinduction verAus postinduction measurements. Asterisk denotes p < 0.01 by Wilcoxon signed rank test.

Discussion Hackground

of

Although measures HRV are now commonly reported in the medical literature, research is ongoing to define the etiology of HRV and the proper interpretation of changes in HRV.“,“J” In most clinical studies, highfrequencv HRV measurements have been interpreted as an index of parasympathetic activity, and low-frequency HRV measurements have been interpreted as an index of sympathetic activity.i,S Pharmacologic blocking studies have established that high-frequency HRV is mediated almost exclusively by ofatroparasympathetic reflexes. w The administration pine essentially abolishes these high-frequency variations.1,2 Furthermore, the known stimulus-response time of the sympathetic system is too sluggish to mediate these more rapid HR changes.’ The primary etiology of highfrequency HRV involves physiologic changes associated with respiration (i.e., the respiratory sinus arrhythmia). Although the precise stimulus for the cyclical increase in HR with inspiration is controversial, postulated stimuli include changes in venous return, pulmonary venous and/or left atria1 stretch, arterial pressure, and lung distention.:” Evidence for the interpretation of highfrequency HRV as an index of parasympathetic tone stems from both experimental intervention studies and observations in disease conditions. Infusion of vasoconstrictors and the recumbent position, both of which are associated with an increase in vagal tone, cause a corresponding increase in high-frequency HRV.S,9’ Conversely, both head-up tilt and the infusion of vasodilators cause a decrease in high-frequency HRV.S High-frequency HRV is reduced in pathophysiologic conditions that have been associated with decreased parasympathetic activity [congestive heart fai1ure,3J hypertension,“”

0

HRV,,

0

HRV,,

06

HR”“, ?

O4 @PM’)

1

i

// /,

0 Fe l”d”CtlOn

ve

LOC

LOC

lnductlo”

Time

Figure 9. Changes in absolute heart rate variability (HRV) power measurements with loss of consciousness (LOC). Note the different y-axis scales for HRV,,, and HRV,,,. (See text for explanation.) Bars indicate SEM. Asterisk denotes measurements that are significantly different from preinduction measurements (p < 0.01 by Wilcoxon signed rank test).

and myocardial infarction (MI)“j]. The most notable of these conditions is that following MI, in which a decrease in certain measures of high-frequency HRV has been repeatedly shown to be a strong, independent predictor of subsequent cardiac morbidity.:“’ Studies using pharmacologic blockade have suggested that low-frequency HRV is mediated jointly by the sympathetic, parasympathetic, and renin-angiotensin sysJ. Clin. Anesrh.,

vol. 4, July/August

1992

271

Original

Contribution

terns. I-$ The etiology of these low-frequency oscillations is less well understood than the etiology of high-frequency HKV. Postulated stimuli include cyclical changes in regional blood flows, thermoregulatory responses, baroreceptor feedback mechanisms, and activity of the reninangiotensin system. Despite the pharmacologic evidence that multiple reflex systems may be involved, low-frequency HRV has been conventionally interpreted in clinical studies as an index of sympathetic. activity.‘,’ Manv studies have confirmed that low-frequency HRV is increased in experimental conditions that cause increased sympathetic activity; examples include head-up tilt, infusion of vasodilators, and moderate exercise.:’ 12,~ Lowfrequency HRV also is increased in pathophysiologic conditions associated with a chronic increase in sympathetic tone (e.g., hypertensionL5’ or MIA?). Prior studies have used both absolute power measurements of HRV and normalized measurements 01. HKV. Absolute power measurements (P.R., HRV,, in the present study) quantitate HRV in defined units derived directly from power spectral analysis. Decreases in absolute HRV power have been correlated with autonomic reflex dysfunction in diabetics,lH quadriplegics,‘7 and patients with congestive heart failure.““:+Absolute HRV powel measurements also have been correlated with changes in baroreflex sensitivity following MI” and have been used in conjunction with measurements of arterial pressure variability to quantitate baroreflex sensitivity.‘2,’ i Normalized measurements of HRV are derived from absolute power measurements of HRV by dividing by a normalizing parameter. This normalizing parameter is some measure of total HRV power (in absolute units). In the present study, %HRV,, [(HRV,,,/HRV,ol) x lOO%] quantitates high-frequency HRV in normalized units. Normalized measurements thus quantitate the fraction of total HRV power contained within defined frequency ranges, irrespective of variations in total HRV power. Normalized measurements of HRV are commonly used in clinical studies to adjust for variations in total HRV among individuals and among study conditions.4,5J2.s4J5 Alternatively, some studies have used other defined ratios of absolute measurements to adjust for these variations (e.g., HRV,,/HRV,.,,).s Many of the original studies of HRV used only absolute power measurements of HRV.‘- 1These early studies suggested that absolute power measurements of HKV correlated well with changes in sympathetic and parasympathetic tone during defined experimental interventions.‘-4 However, recent studies have documented that under some circumstances of sympathetic activation (CR., exercise or mental stress), absolute power measurements of’low-frequency HRV were reduced.’ This paradoxical reduction in low-frequency HKV (expressed in absolute units) during certain circumstances of increased sympathetic tone apparently stems from the reduction in total HRV during the experimental intervention. However, examination of the change in the frequency distribution of HRV demonstrated the expected shift toward a greater percentage of HRV power in the low-frequency range. This shift in frequency distribution was reflected 272

J. Clin. Anesth.. vol. 4, .July/August 1992

bv an increase in normalized measurement,5 of lowfrequency HKV (in accordance with increased sympathetic tone). ‘I‘hese later studies suggest that under conditions of sympathetic activation when total HKV is reduced, changes in absolute power measurements (of low-frequency HRV may not correlate with changes in sympathetic tone; however, changes in normalized measurements of HRV still appear to reflect changes in sympathetic-parasympathetic balance.-’

Our results document very clearly that tiif’f&-ent me+ thetic inductions do have different effects on HKV. ‘l‘hr thiopental sodium induction technique caused -LII 89% t 2% reduction in total HRV ZY~SZL.\ a 5X’% at 13% I-C‘duction caused by a comparable etomidate technique. Whereas the thiopental sodium induction caused a significant decrease in %HRV,,,, the etomidate induction had no significant effect on normalized HRV measurements. ‘I’he greater depressant effect ot the thiopental sodium rechnique on total HRV is consistent with the relative eflec-ts of thiopental sodium and etomidatr OII cardiovascular reflexes. The greater depressant effect of thiopental sodium on baroreflexes has been documented by two studies that made direct comparisons between lhese two drugs,‘” Z2 as well as by other studies 01 t,arbitur;tres:‘~~“’ and etonlidate.4U The greater depressant effect of thiopental sodium on HRV,,, IWI.S%IS HKV, (,, suggesting a greater depressant effect on parasympathetic 7~3~s sympathetic reflexes, also is consistent with prior studies of‘barbiturates.rKml” These relative effects of‘thiopental sodium on HRV,,, and HRV,,, caused a significant reduction in ‘%HRV,,,. ‘I‘his reduction in SHRV,,, also is consistent with the vagolytic action of barbiturates and suggests that normalized measurements of HRV mav reflect changes in sympathetic-parasympathetic balance. even when total HRV is reduced in anesthetized patients. These relative effects of thiopental sodium and etomidate on tfRV suggest that anesthetic effects on HRL’ are due in part to effects on cardiovascular reflexes. It should be noted, however, that the observed reductions in absolute power measurements of HRV with both techniques were of’ much greater magnitude than reported ef’fects on baroretlexes.“,‘~ J’l‘his is particularly evident technique. Whereas etomidate has with I he etomidate minimal effects on baroreflexes, etomidate induction caused a 58% % I 3c% reduction in total HRV. l‘hesr findings suggest chat mechanisms other than baroreflex depression may be involved with the reduction in HRV caused by anesthetics. Possible additional mechanisms for the depressant el’fects of anest hetics on HKV include anesthetic effects on centrally mediated reflexes and cognitive function. Studies in brain-injured patients and quadriplegics have shown that central integration of cardiorespiratory reflexes is required for the genesis of HRV. 16,1iGeneral anesthesia mav impair the central integration and modulation of these reflexes (at either central or ganglionic sites). In

Anesthetic effects on heart rate variabilitl;: Latson et al.

patients with episodic increases of intracranial pressure, changes in HRV paralleled changes in cognitive function (level of consciousness, orientation).‘5 This suggests that cerebral function, and perhaps consciousness per se, may play an important role in HRV. This reduction in HRV caused by anesthetics also may be related to reported reductions in central autonomic outflow recorded by other techniques. Using decerebrate rabbits, Hughes and MacKenzie” found that while etomidate had no effect on baroreflexes, it did cause a significant reduction in preganglionic sympathetic outflow (similar to that observed with thiopentone, ketamine, and althesin). More recent studies in humans have shown reductions in muscle sympathetic nerve activity with induction of anesthesia by thiopental sodium ,4*etomidate,22 and propofol.dg As early as 1965, McCrady et al6 suggested that changes in the respiratory sinus arrhythmia (RSA; analogous to high-frequency HRV) might provide “a sensitive and quantitative means of monitoring levels of anesthesia.” Results from this early dog study suggested that the effects of sodium pentobarbital and halothane on RSA were similar for comparable stages of anesthesia. Donreported a 70% decrease in RSA chin et al.’ subsequently (measured specifically as the variance of the HR pattern in the frequency band of respiration) during isofluraneN,O anesthesia; RSA then returned to near control values in the recovery room. These studies were later extended by Pfeifer et uL.,~ who demonstrated similar dose-related RSA reductions in patients anesthetized with enflurane and N,O and suggested that RSA might provide a useful “index of anesthetic depth.” Latson et al.’ also examined the use of HRV as a monitor of anesthetic depth and documented that increases in high-frequency HRV usually precede intraoperative changes in lower esophageal contractility. The prospective comparison of two anesthetic techniques in the present study adds important information regarding the use of HRV as a monitor of anesthetic depth. In contrast to McCrady et al.‘s study: the present study documents that different anesthetics can have different effects on HRV. This result suggests that if analysis of HRV is used to monitor anesthetic depth, different criteria for assessing depth may be required for different anesthetic techniques.

Study 2 The primary reason for studying sufentanil was its purported effects on vagal activity. Similar to fentanyl, sufentanil is believed to increase vagal tone.44 Evidence of this action of potent opioids comes from studies using pharmacologic blockade, vagotomy,*3 brain stem transection,Z4 and direct recordings from medullary neurons.‘j The results from prior studies correlating changes in absolute measurements of high-frequency HRV with vagal activity”2.4j.46 would suggest that the increase in vagal activity caused by sufentanil should have caused an increase in HRV,,. To the contrary, sufentanil induction caused a 69% 2 12% reduction in HRV,,. This finding

is similar to that previously reported by Komatsu et al.” using a fentanyl induction technique. These investigators reported that induction of anesthesia with fentanyl 50 to 70 pg/kg, diazepam 0.15 to 0.3 mgikg, and pancuronium 0.1 mg/kg caused a 71% reduction in absolute power measurements of high-frequency HRV. In the present study, we avoided the concomitant administration of benzodiazepines and used a muscle relaxant that has no significant chronotropic or sympathomimetic properties. Despite these differences, the results are very similar. These studies document that changes in absolute

power measurements of high-frequency HRV are not a valid measure of changes in vagal tone during potent opioid induction. This failure of HRV,, to correlate with vagal activity might result from a decrease in baroreflex sensitivity. Although the effects of sufentanil on baroreflex sensitivity have not been reported, a prior study with fentanyl suggested that parasympathetic baroreflex control remained intact after fentanyl administration (10 and 12.5 pgikg, in combination with diazepam and N&J).47 Alternatively, this failure of absolute power measurements of HRV to correlate with changes in autonomic tone may be similar to that described in awake subjects during certain experimental interventions (e.g., mental stress) accompanied by a decrease in total HRV. Under such circumstances, absolute power measurements of lowfrequency HRV may not correlate with changes in sympathetic tone.4 In our experiments, there was a dramatic reduction in total HRV between the awake and anesthetized conditions ( - 83% ? 6%). We believe this dramatic reduction in total HRV is the most likely explanation for the failure of HRV,, to correlate with changes in vagal tone during sufentanil induction. The etiology for this global reduction in HRV observed during sufentanil induction is beyond the scope of this study. This reduction in total HRV, in the presence of pharmacologic parasympathetic activation, may represent an effect similar to the reduction in total HRV observed during some conditions of sympathetic activation.’ Alternatively, this reduction in total HRV may be an effect related to general anesthesia. Observations supporting the latter explanation include the following: (1) all prior studies with general anesthetics have also demonstrated a reduction in HRV between the awake and anesthetized states; (2) our trended measurements of HRV showed a close temporal relationship between this global reduction in HRV and loss of consciousness. In contrast to the effects of sufentanil on absolute power measurements of HRV, the effects of sufentanil on normalized measurements of HRV are consistent with a relative increase in parasympathetic activity. Although sufentanil reduced HRV,, by 69% -C 12%, it caused an almost twofold increase in %HRV,, (from 24% -t 4% to 41% 2 6%, p < 0.01). This increase in %HRV,, reflects the greater reduction in HRV,,, (- 88% + 5%) *Komatsu T, Kimura T, Sanchala V, Sbibutani K, Lees DE: Evaluation of spectrum analysis of heart rate variations during anesthesia [Abstract]. Anesthesiology 1986;6.5:.4 139. J. Clin. Anesth., vol. 4, .July/August

1992

273

’

Original Contributions

relative to HKV,,, (- 69’3 ? 12%) following sufkntattil administration. A similar finding was reported in the study by Komatsu et nl., * in which induction of‘anesthesia with fentanyl, diazepam, and pancuronium increased the ratio of high-frequency to low-frequency HKV from 45% to 62%. These directionally opposite changes in HKV,,, and %HKV,,, result from the normalizing procedure ittvolved in the calculation of’ the latter parameter. ‘fhis normalizing procedure adjusts for the reduction in total HKV power that we observed following patients’ loss of consciousness. Similar to results in awake subjects’ (and similar to our results with thiopental sodium), these results suggest that normalized measurements of’HKV ma\’ still provide a qualitative index of changes in sympathetic-parasympathetic balance, even when total HKV ih reduced.

Study LimitaIions ‘l‘he use of’ancillary drugs (premedicants, lidocaine. and N#) does impose limitations on comparisons between out- results and those of’ prior studies investigating the autonotnic effects of’these drugs. In Study 1, we wished to make comparisons using standard equivalent doses of’ thiopental sodium and etomidate. We fintnd that these drugs by themselves sometimes resulted in less than satisfactory inductions in these otherwise unpremeditated young patients. The ancillary drugs administered IO et)sure a smooth induction (sufentanil premeditation, lidocaine, and N#) were given to both groups in identical fashion. We did not observe any significant ef’fect of’this small dose of’ sufentanil on the trended measurements of’ HKV, but a possible synergistic effect with other anesthetics administered camlot be excluded. Kecettt studies suggest that N,O does not af’f’ect baroreflex augmentation of muscle sympathetic nerve activity Is or vasoconstrictor responses:” but may influence baroreflex control of’ HK.“* Other studies suggest that N,O ma) increase central sympathetic outflow.i’3,‘X Although pt-evious studies suggested that lidocaine does not have an) significant autonomic effects,“” a recent investigation by Ebert et aI.” documented attenuation of sytnpathetic responses to autonomic stress. The possible effects of‘N,O and/or lidocaine on HKV have not been investigated. In Study 2, we felt that it would be unethical to withhold premeditation from these patients about to undergo cardiac surgery, in whom increased anxiety may provoke myocardial ischemia. Midazolam premeditation was selected because of its potent anxiolytic effect. The t,yte of premeditation administered prior to a potent optotd induction may influence the hemodynamic and autonomic responses to induction.52 The effects of midazolam on HKV have not been reported. l‘here were considerable differences in the preinduction HKV measurements of patients in Study 1 compared with those in Study 2 (Tahlrs I and 2). These *I(o~r~atsu I’, I(imt~ra .r, Sanchala

V, Shibutani K. Lres 1)E: tvaoration of spectrum analysis of heart I-ate variations during a~\thesia (Ahstrarr]. A~est/w~iolo~gv 1986;65:A139.

274

J. Clin.

Anesth., vol. 4, July/August

1992

tlil’fkrettccs reflect the efftcts of age and other ttiedic.al conditions in Study 2 patients. Decreases in higltfieqttettc-y HK\’ have been associated with increasing age,” coronary artery disease,” hypertension, \’ cortgestive heat-t failure.“’ recent, MI,‘” and diabetes.‘” Etiologies f’or these changes in HKV include reduced baroreceptotsensitivity, induced alterations in ittttottot~iic‘ tone (P.S.. increased sympathetic activity and/or decreased vagal tom in cardiovascular disease), and autonomic. neuropathv. l’his population of‘ patients undergoing heart surg& was selected for Study 2 because the described sufentattil induction technique is c-ommotily used for these patients (whereas it is rarely used in ASA physical status I patients). Whether these dif’f’erences between stud) popttlatiotts ma)’ have affected our results is unknown. (:omparecl with the thiopental sodiutn induc~tiott, the rtomidate induction had much more \ ariable ettects on IIKV. l’his variability is illustrated by the tnuch largelstandard errors fi)r the etomidate data in F‘~~IIIP F. ‘l‘ltis ittcrcasctt variability in the respottse to etomidatc tits\ have litnitctl ourabilily (relative to the thiopental sodium group) to detect dif&rettces in the cfl’ects 01’etotttidatc on HKV,,, and IHKV,,,. Additional sfudies in it Iat-get, number of’ patients would be requiret1 to adtires\ this issue. .I.ltis irtcreastd \,ariability also ttta!’ have limited our abilit\~ to detec.1 a possible dif‘frrettce itt the rl‘fec~ts of tltic,pe;ttal sodium and etomidate on HKC’,,,.

Conclusions Ek.11of these induction it11 f’recptmc~

techniques caused tedttctions itt with the of HKV consistent examining changes in HKV tlur-

cotrtpmtm~s

results 01’prior studies ing general anest besia. However, there wrre considerable dift’et-ettcrs in the ett’ects of’the diflerrnt techniques. Such differences indicate that drug-specific criteria ma) be required when using HKV measuremettts fi)r monitoring depth of anesthesia. The relative effects ol’ etomidatr itttd thiopental sodium Ott HKV, AS well :IS the tttct‘ease tn “cHKV,,, with suf’entanil. provide eviderrw that changes in HKV caused by anesthetics at-e t&ted in part to known effects on autonomic. reflexes. However, bot II the large reduction in total HK\’ c-aused I)\ the etomidate induction and the reduction in HKV,,, caused b) suf’etttanil suggest that additiottal pf’f’ect.s of general atlesthesia (u.g., impaired consciousness) also nta! be involved in these changes. References

Anesthetic effects on heart rate variability: Latson et al. 4. Malliani A, Pagani M, Lombardi F, Cerutti neural regulation explored in the frequency

S: Cardiovascular domain. Ctrcula-

tion 1991;84:482-92. 5. Furlan R, Guzzetti S, Crivellaro W, et al: Continuous 24.hour assessment of the neural regulation of systemic arterial pressure and RR variabilities in ambulant sukjects. Czrculatzon 1990;81:537-47. 6, McCrady JD,Vallbona C, Hoff HE: The effect of preanesthetic and anesthetic agents on the respiration-heart rate response of dogs. AmJ Vet Rer 1965;26:710-6. 7. Donchin Y, Feld JM, Porges SW: Respiratory sinus arrhythmia during recovery from isoflurane-nitrous oxide anesthesia. Anstb Atu~lg 1985:64:81 l-5. 8. Pfeiter BL, Sernaker HL, Porges SW: Respiratory sinus arrhythmia: an index of anesthetic depth? [Abstract]. Ar~esthAnal, 1988;67:S 170. 9. Latson 1‘. Martin D, Isaac P: lntraoperative changes in respiratory sinus arrhythmia: ? an indicator 01. anesthetic- depth. [Abstract]. Anr.cthe,\iol~!yy 1990;73:A520. 0: Spectral analysis of heart IO. lshikawa -r, Kimura T, Kemmotsu rate in sevoflul-aneinitrous oxide anesthesia [Abstract]. Arxthr\ioto
25.

Ldubie

M, Schmitt

H: Action

fentanyl. on the nucleas tractus biguus cardiovascular neurons. 26.

27.

28. 29. 30.

3 1.

32.

33.

of the morphinometic

agent,

solitarii and the nucleus amEur _/ f’harmnco~ 1980;67:

403- 12. Berger RD, Akselrod algorithm for spectral

S, Gordon D, Cohen RJ: An efficient analysis of heart rate variability. IEEE Tmn( Bzomrd Eng 1986; BME-33:900-4. Albrecht P, (lohen RJ: Estimation of heart rate power spectrum bands from real-world data: dealing with ectopic beats and noisy data. (,‘omput Cardiol 1989: 17:sI l-4. Brigham EO: The Fast Fourier Transform. Englewood Cliffs, NJ: Prentice-Hall, 1974: 140-6. <:hen
34.

Guzzetti S, Piccaluga E, Casati R, et al: Sympathetic predominance in essential hypertension: a study employing spectral analysis of heart rate variability. J Hypertew 1988;6: 711-7. 35. Lombardi F, Sandrone G, Pernpruner S, et al: Heart rate variability as an index of sympathovagal interaction after acute myocardial infarction. Am.1 Cardiol 1987;60: 1239-45. 36. Kteiger RE. Miller .JP, Krone RJ, Bigger JR Jr, Multicenter Postint>rction Research Group: The independence of cycle length val-iability and exercise testing on predicting mortality of patients surviving acute myocardial intarction. Am,] Cardiol 1990;65:4081 1. 37.

3X.

39.

40.

41.

42.

43.

44.

Rimoldi 0, Pierini S, Ferrari A, Cerutti S, Qagani M, Malliani A: Analysis of short-term oscillations of R-R and arterial pressure in conscious dogs. Am J Phy,vol 1990;25H:H967-76. Mac-Kenzie JE, Mcgrath JC, Tetrault JP. Millar RA: The effects of althesin and thiopentone on sympathetic and baroreflex ac-tivitv. &n Ar~esfh Socj 1976:23:252-62. Vatner SF, Franklin D, Braunwald E: Effects of anesthesia and sleep on circulatory response to carotid sinus nerve stimulation. Ani .I I’/L?\/Ol 197 1;220: 1249-55. Bristow JD, Prys-Roberts (:. Fisher A. Pickering TG, Sleight P: Effects of anesthesia on baroreflex control of heart rate in man. .4rtrll~e,\zolo,~ 1969;31 :422-X. Hughes RL, MacKenzie JE: An investigation of the centrally and peripherally mediated cardiovascular effects of etomidate in the rabbit. BP-,JAnaesth 1978;50: 101-7. Ebrrt 111, Kanitz DD, Kampine JP: Inhibition of sympathetic neural outflow during thiopentdl anesthesia in humans. Anesth Ar& 1990;71:319-26. Sellgren J, Ponten J, Wallin (;B: Percutaneous recording of muscle nerve sympathetic activity during propofol, nitrous oxide, and isoflurane anesthesia in humans. Anesthesiology 1990:73:20-7. <;ravlee GP, Ramsey FM, Roy KC, Anger1 KC, Rogers AT, Pauca AF: Rapid administration of a narcotic and neuromuscular blocker: a hemodynamic comparison of fentanyl, sufen-

J. Clin.

Anesth.,

vol. 4, July/August

1992

275

Original Contributions tanil, pancut-onium, aud vecuronium. Anoth ilnulg 19X8:67: 39-47. 45. Katona I’(;, Jih F: Respiratory sinus arrhythmia: noninvasive measure of parasympathetic cardiac control. ,/ App/ Phytiol 1975;39:801-5. 46. Eckberg 111,: Human sinus arrhythmia as an index of vagal cardiac outflow. ,/ Appl Phyzol 1983;54:961-6. 47. Kotrly KJ, Ebert TJ, Vucins EJ, Roerig DL, Staduicka A, Kampine JP: Effects of fentanyl-diazepam-nitrous oxide anaesthesia on arterial baroreflex control of heart rate in man. Hr ,J Anaesth 1986;58:406-14.

4X. Ebert ‘FJ: Differential

effects of‘ nitrous oxide on haroreflcx

control of‘heart rate and peripheral sympathetir nerve activit\ in humans. Anrsthedogy 1990;72: 1G22. 49. Ebert T,J, Kotrly KJ, Madsen KE, Bernstein JS, Kampine JP: Fentanyl-diazepam anesthesia with or without N,O does not attenuate cardiopulmonary baroreflex-mediated vasoconstrit for responses to controlled hypovolemia in humans. And/r A,,n(q 1988;67:.548-54.

Since the publication of the classic and authoritative book of Gillespie, a great change has taken place, and this change has largely been due to this book as well as to other authoritative articles written by anesthesiologists. The complication of gross laryngeal and adjacent tissue trauma has been eliminated by the development of skill in laryngeal technic on the part of the anesthesiologists. The other laryngeal complications recorded in the literature are contact ulcer granuloma, acute edema of the larynx requiring tracheotomy, and asphyxia.

Contact

Ulcer Granuloma

In the early reports of laryngeal granuloma as a sequela of endotracheal anesthesia it was regarded as a newly discovered lesion. Farrior (1942) noted that the lesion in these cases is a typical contact ulcer with granuloma formation, described and illustrated by Jackson. New and Devine (1949) made the same observation; they collected from the literature 9 cases of contact ulcer granuloma following endotracheal anesthesia and added 9 cases from their own experience. All of t.he contact ulcers so far reported in the literature as sequent to endotracheal anesthesia were in the granuloma state, hence they were all cases of delayed diagnosis. It takes time for a granuloma to develop. is most commonly superimposed on As stated by Jackson and Jackson, “Nonspecific granuloma a contact ulcer.” Charles M. Norris stated that at the Temple University Hospital during the five y-ear period 1948 through 1952, 48 cases of contact ulcer granuloma were recorded, of which b (12.6 pel cent) followed endotracheal anesthesia. In 3 of the 6 cases, the granuloma occurred while the patients were in the hospital (2 after thyroidectomy and one after cervical spine operation for dislocated intervertebral disc). The other 3 patients had had operations elsewhere under endotracheal anesthesia. Contact ulcer of the larynx was first described in 1928 by Chevalier Jackson, on the basis 01 248 Cases observed in 40 years. In 1935 Jackson and Jackson reported 45 additional cases, making a total of 293 cases. In none of these cases was the laryngeal condition a sequela of endotracheal anesthesia or of endolaryngeal instrumentation. Sixteen colored illustrations from oil-color paintings from life show that in all but 2 of the cases illustrated the contact ulcer was in the granuloma stage when first examined. Jackson C: Contact ulcer granuloma anesthesia. Anesthesiology 1953; 14:425-6.

From

276

J. Clin. Anesth.,

vol. 4, July/August

1992

and other laryngeal complications Reprinted with permission.

of endotracheal