Effects of Intrathecal Morphine on Transcranial

Electric Motor-Evoked Potentials in Adolescents

Undergoing Posterior Spinal Fusion
Paul A. Stricker, MD,*† Anthony K. Sestokas, PhD, D. ABNM, F. ASNM,‡
Daniel Schwartz, PhD, D. ABNM, F. ASNM,‡ Vidya Bhalodia, MS, D. ABNM,‡ Ankit Pahwa, MS,‡
John P. Dormans, MD, FACS,§储 Lia L. LaBrant, BA,*† Beverly J. Burgess, BA,*† F. Wickham Kraemer, MD,*†
and Arjunan Ganesh, MBBS*†

BACKGROUND: Intrathecal morphine (ITM) provides effective analgesia after posterior spinal
fusion (PSF). Although most anesthetic drugs have well-characterized effects on evoked
potentials, there is little data on the effects of ITM on transcranial electric motor-evoked
potentials (tceMEPs). We performed this study to assess the effects of ITM on tceMEPs in the
first 30 minutes after administration. We hypothesized that administration of ITM in doses
currently used at our institution would not significantly affect mean tceMEP amplitudes and
latencies of an ITM study group relative to control patients who did not receive the drug.
METHODS: tceMEPs were recorded before ITM injection and 5, 10, 20, and 30 minutes after
injection in 14 subjects ages 11 through 18 years undergoing PSF. These recordings were
compared to an age-matched control group undergoing PSF in which ITM was not injected. The
effects of ITM on tceMEP amplitude and latency were compared between the 2 groups.
RESULTS: Fourteen subjects were enrolled in the ITM group and 16 served as controls. There
were no significant differences in the baseline mean response amplitudes of the 2 groups for any
of the 8 muscles studied. Mean response amplitudes over the 30-minute posttreatment period
in the ITM group did not differ significantly from those of the control subjects. Average response
amplitudes collapsed across all muscles for each subject were not significantly different during
the baseline period (95% CI ⫽ ⫺38% to 45%; P ⫽ 0.783), nor were they significantly different
between the 2 groups during the posttreatment period (95% CI ⫽ ⫺30% to 78%; P ⫽ 0.640).
There also were no significant differences in the mean response latencies of the 2 groups in
either the baseline or posttreatment periods. Average response latencies collapsed across all
muscles for each subject were 4% larger for the ITM group than for controls during the baseline
period (95% CI ⫽ ⫺5% to 13%; P ⫽ 0.377), and 3% larger for the ITM group than for controls
during the posttreatment period (95% CI ⫽ ⫺4% to 12%; P ⫽ 0.359).
CONCLUSIONS: Administration of ITM in doses currently used at our institution did not cause
more than a 70% attenuation of mean tceMEP amplitudes or latency changes of an ITM study
group relative to control subjects during the 30-minute period after injection. Further studies are
required to determine if there are delayed effects after this initial time period. (Anesth Analg
2012;115:160 –9)

P atients undergoing posterior spinal fusion (PSF) for

idiopathic scoliosis are at risk for neurologic injury
during surgical correction. Intraoperative monitor-
ing of evoked potentials is performed to identify impend-
ing neurologic injury. Transcranial electric motor-evoked
From the *Department of Anesthesiology and Critical Care Medicine, The
Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania; †Depart-
ment of Anesthesiology and Critical Care Medicine, Perelman School of
Medicine at the University of Pennsylvania, Philadelphia, Pennsylvania;
‡Surgical Monitoring Associates, Springfield, Pennsylvania; §Department of
Orthopedic Surgery, Children’s Hospital of Philadelphia, Philadelphia,
Pennsylvania; and 储Department of Orthopedic Surgery, Perelman School of
Medicine at the University of Pennsylvania, Philadelphia, Pennsylvania.
Accepted for publication February 2, 2012.
This study was funded by the Department of Anesthesiology and Critical
Care Medicine of The Children’s Hospital of Philadelphia and the University
of Pennsylvania School of Medicine.
Conflicts of interest: See Disclosures at the end of the article.
Reprints will not be available from the authors.
Address correspondence to Paul A. Stricker, MD, Department of Anesthe-
siology and Critical Care Medicine, The Children’s Hospital of Philadelphia,
34th Street and Civic Center Boulevard, Philadelphia, PA 19104-4399.
Address e-mail to strickerp@email.chop.edu.
Copyright © 2012 International Anesthesia Research Society
DOI: 10.1213/ANE.0b013e31824e5d86
potentials (tceMEPs) are elicited using multipulse electrical
stimulation applied through the scalp and skull to intact
cortex, thereby exciting pyramidal cells in the motor cortex.
Pyramidal cell axons converge to form the corticospinal
tract, which terminates in spinal cord ␣-motoneuron pools
that innervate efferent spinal motor nerve roots. These
roots in turn terminate via peripheral nerves at muscular
motor endplates. Compound muscle action potentials are
then recorded at the target muscle. During PSF, tceMEPs
are used to monitor the integrity of the lateral corticospinal
tract and nerve root motor pathways.
Most commonly used anesthetics have well-characterized
effects on evoked potentials.1–7 Although the effects of intra-
thecal morphine (ITM) on somatosensory-evoked poten-
tials have been described,8 there is little data on its
effects on motor-evoked potentials. Pronounced effects
of ITM on motor function have been demonstrated in
animal models;9 –11 therefore, this study sought to inves-
tigate early onset effects of ITM on tceMEP during PSF.
Our hypothesis was that administration of ITM to a
group of patients in doses currently used at our institu-
tion would have no significant effects on mean tceMEP

160 www.anesthesia-analgesia.org July 2012 • Volume 115 • Number 1

amplitudes and latencies during the initial time period
after injection.
METHODS
This prospective observational study was approved by the
IRB of The Children’s Hospital of Philadelphia. Written
informed consent was waived by the IRB, and verbal
informed consent and assent were obtained for all subjects
enrolled. Subjects aged 11 to 18 years undergoing PSF were
eligible for enrollment. Study group assignment was deter-
mined by the caudal extent of the planned procedure.
Subjects were enrolled in the ITM group if the planned
surgical procedure extended to or below the second lumbar
vertebral level and administration of ITM was planned.
Subjects were enrolled in the control group if the planned
surgical procedure did not extend below the second lumbar
vertebral level and administration of ITM was not planned.
Subjects were excluded if they received neuromuscular
blocking drugs or inhaled anesthetic drugs after the start of
surgery because of the profound effects that these drugs
can have on tceMEPs. Subjects who developed clinically
significant changes in tceMEPs intraoperatively before on-
set of the study period, as well as those who developed
intraoperative hemodynamic instability requiring a con-
tinuous vasoactive drug infusion, were also excluded.
After induction of anesthesia (either IV or inhaled)
and insertion of an IV catheter, anesthesia was main-
tained using a total IV technique with infusions of
propofol, remifentanil, and ketamine. No neuromuscular
blocking drugs or inhalation anesthetics were given for
at least 3 hours before the collection of study data.
Ketamine infusion rates were kept constant during the
study interval and boluses were not administered. Intra-
thecal morphine injections were performed by the sur-
geon under direct visualization intraoperatively.
tceMEPs were monitored throughout the surgical course
as part of a multimodality monitoring strategy that also
included ulnar and posterior tibial nerve somatosensory-
evoked potentials for assessment of conduction along pe-
ripheral and central ascending sensory pathways, train-of-4
peripheral nerve stimulation to evaluate conduction across
the neuromuscular junction, spontaneous and stimulated
electromyography during placement of pedicle screws, and
2-channel electroencephalography for facilitated assess-
ment of anesthetic depth, as described elsewhere.12 Tce-
MEPs were triggered by delivering brief electrical pulse
trains between subdermal needle electrodes placed at scalp
locations C1 and C2 (International 10 to 20 System) over-
lying primary motor cortex. Pulse train parameters were
adjusted at the start of the procedure to maximize response
amplitudes. The number of pulses typically ranged be-
tween 3 and 7, with a 1 to 5 ms interpulse interval. Stimulus
intensity was varied between 300 and 700 volts.
Transcranial stimuli were delivered using either a Digi-
timer D185 multipulse stimulator (Digitimer, Welwyn Garden
City, UK) or special-purpose internal stimulators inherent to
the neuromonitoring systems used for this series. Collection
of neuromonitoring data was accomplished using Axon
Epoch/Eclipse (Axon Systems, Hauppauge, NY), Nicolet En-
deavor (Viasys Healthcare, Madison, WI) or Cadwell Cascade
July 2012 • Volume 115 • Number 1
(Cadwell Laboratories, Kennewick, WA) monitoring systems,
having a minimum of 16 recording channels.
After uneventful correction of the spinal deformity,
preservative-free morphine 0.5 mg/mL (3 to 16 ␮g/kg,
Astramorph, APP Pharmaceuticals, Schaumburg, IL) was
slowly injected by the surgeon after aspiration of cerebro-
spinal fluid using a 25-gauge Whitacre needle via the L2 to
L3 interspace. Baseline tceMEPs were recorded just before
administration of ITM for each subject in the ITM group
and at the start of surgical wound closure for each
patient in the control group. These responses were
designated as the pretreatment baselines for both groups.
Responses were recorded from bilateral first dorsal in-
terosseous muscles in the upper extremities, and from
bilateral quadriceps, tibialis anterior, and abductor hal-
lucis muscles in the lower extremities. TceMEPs subse-
quently were recorded from the same muscles 5, 10, 20,
and 30 minutes after baseline recordings. Invasive arte-
rial blood pressure measurements were recorded at time
points correlating to tceMEP measurements.
Data Analysis
Amplitudes and latencies of tceMEP waveforms were mea-
sured by an independent member of the research team who
was blinded to both the procedure and study group
assignment. Response amplitude was defined as the differ-
ence between the highest peak and lowest trough in each
waveform (peak-to-trough amplitude). Response latency
was defined as the elapsed time between stimulus onset
and the take-off point from the resting potential of each
waveform.
Natural log transformations of amplitude and latency data
were applied in an effort to improve symmetry/normality
and reduce heteroscedasticity of distributions before statistical
analysis. For purposes of graphical display, antilogs of the
transformed distribution means and 95% CI were used to
summarize central tendency (geometric mean) and response
dispersion, respectively, for each group.
The primary null hypothesis of the study was that
administration of ITM would not cause a clinically signifi-
cant attenuation of mean tceMEP amplitudes (i.e., that any
amplitude attenuation would be ⬍70%) or prolong mean
latencies of ITM study subjects relative to age-matched
controls who did not receive the drug. We estimated that
inclusion of 15 patients per group would provide 80%
statistical power (with an ␣ level ⫽ 0.05) to identify a
clinically significant effect size of 70% amplitude attenua-
tion after administration of ITM. This effect-size estimate
was based on clinical criteria for significant tceMEP ampli-
tude change during pediatric spine surgery under anes-
thetic conditions similar to those of the present study.12,13
Use of a clinical alert criterion of at least 70% amplitude
attenuation considers the known inherent variability of
tceMEPs and reduces the likelihood of false-positive results
associated with less stringent criteria.12
Pretreatment baseline response amplitudes of ITM and
control patients were compared using a 2 ⫻ 4 ⫻ 2 mixed
model ANOVA (SYSTAT 13, Systat Software Inc., Chicago,
IL) to identify any inherent differences in response size for
the 2 treatment groups across 4 different muscles on both
sides of the body. This was followed by a similar analysis of
www.anesthesia-analgesia.org 161
Intrathecal Morphine Effects on Motor-Evoked Potentials

Table 1. Demographics Table 2. Mean Arterial Blood Pressure (mmHG)

Intrathecal at Transcranial Electric Motor-Evoked Potential
Demographics Morphine Control P Measurement Time Points in Control and
Age 14 ⫾ 2 y 14 ⫾ 2 y 0.98 Intrathecal Morphine (ITM) Study Groups
Weight (kg) 55.1 ⫾ 12.8 55.5 ⫾ 13.7 0.93 Subject # Study Group Baseline 5 min 10 min 20 min 30 min
Gender 1 Control 67 72 61 60 51
Male/female 2/12 5/11 4 Control 92 97 99 70 89
Number of vertebral levels 12 ⫾ 1 12 ⫾ 1 0.79 5 Control 64 68 66 65 73
operated on 6 Control 70 66 66 68 68
Estimated blood loss (mL) 784 ⫾ 291 989 ⫾ 599 0.27 7 Control 75 68 68 76 80
9 Control 56 58 61 65 66
10 Control 67 65 68 67 72
mean response amplitude over the 30-minute posttreat- 11 Control 66 72 71 62 85
ment period. For each muscle, mean response amplitude 12 Control 68 60 73 68 66
was defined as the arithmetic average of the 4 response 13 Control 63 72 79 74 77
14 Control 50 71 70 61 63
amplitudes measured 5, 10, 20, and 30 minutes after 15 Control 75 81 67 61 68
pretreatment baseline recordings, respectively. To address 17 Control 66 79 77 70 64
possible effects of chance baseline differences between the 19 Control 67 72 65 67 70
ITM and control groups, the second analysis was supple- 20 Control 58 79 67 65 61
21 Control 83 82 83 93 80
mented by ANCOVA using baseline response amplitude as
2 ITM 59 66 62 69 67
the covariate. Effects of treatment group also were investi- 3 ITM 70 73 95 97 72
gated using relative measures of tceMEP amplitude for 8 ITM 66 57 61 70 65
each muscle. Relative measures were obtained by calculat- 16 ITM 73 71 63 71 70
ing ratios of the 4 posttreatment response amplitudes to 18 ITM 59 57 63 66 64
22 ITM 69 72 72 64 65
that of baseline for each subject. The median of these 4 23 ITM 64 91 86 81 82
ratios was then used as the dependent measure for 24 ITM 83 77 74 73 63
ANOVA. Analyses similar to those described above were 25 ITM 66 71 57 66 64
performed on response onset latencies to compare temporal 26 ITM 78 67 60 64 74
27 ITM 63 61 59 64 60
characteristics of tceMEPs for ITM and control subjects.
28 ITM 69 68 65 69 69
ANOVA was supplemented by post hoc t-test comparisons 29 ITM 58 57 52 65 48
of select responses (2-sample assuming unequal variances), 30 ITM 54 56 68 66 60
as described in the Results section. The effects of multiple
t-test comparisons were addressed with Bonferroni correc-
tion to retain a family-wise type-I error rate of 0.05.
Reported 95% CIs for post hoc t-tests include correction for Table 3. Skewness and Kurtosis of Baseline
the effects of multiple comparisons. Transcranial Electric Motor-Evoked Potential
Amplitude and Latency Distributions, Evaluated
Before and After Natural Log Transformation
RESULTS of the Data
Thirty-one subjects were enrolled in the study. One ITM
Before transformation After transformation
group subject did not receive the drug because the surgical
procedure did not ultimately extend below the second Measure Skewness Kurtosis Skewness Kurtosis
lumbar vertebral level; therefore this subject was moved to Amplitude 1.73 3.55 ⫺0.71 0.76
Latency 0.99 1.34 0.69 0.57
the control group. At the time of analysis, the recordings of
1 of the ITM subjects could not be recovered. This yielded The table shows median values of skewness and kurtosis for response
distributions of eight muscles.
14 subjects in the ITM group and 16 subjects in the control
group. Demographic data are presented in Table 1. Dosages
of ITM were at the discretion of the attending anesthesiolo- group responses in the posttreatment period. The pretreat-
gist, with a median of 7 ␮g/kg (range 3 to 16 ␮g/kg). ment baseline data of ITM and control patients were pooled
Hypotension, defined as a decrease of mean arterial blood for this initial characterization. In general, the distri-
pressure below 60 mm Hg that is more than a 15% decrease butions of amplitude and latency measurements were
from baseline was observed at 2 of 120 time points at which positively skewed and had positive excess kurtosis rela-
tceMEPs were recorded (Table 2). By protocol, tceMEPs tive to the normal distribution. Natural log transforma-
were to be recorded from 8 muscles at 5 specified times for tions of these measurements decreased the amount of
each of the 30 subjects enrolled in the study, yielding a total skewing and kurtosis, as summarized in Table 3 and
of 1200 responses in the study database. Five subjects had facilitated data normalization.
incomplete datasets due to deviations from study protocol The Shapiro-Wilk test was used to assess normality of
or presence of artifact in some recordings, resulting in loss response amplitude and latency distributions for each of 8
of 33 data points and an overall missing data rate of 2.75% muscles, as summarized in Tables 4 and 5, respectively.
for the study. Before transformation of data, the null hypothesis that each
The distributions of baseline tceMEP amplitudes and la- sample of 30 observations came from a normally distributed
tencies for 8 muscles were examined graphically and quanti- population was rejected for all 8 amplitude distributions and
tatively before statistical comparison of ITM and control 4 of 8 latency distributions. After log transformation, the null

162 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

 Table 4. Results of Shapiro-Wilk Tests for
Normality of Baseline Transcranial Electric Motor-
Evoked Potential Amplitude Distributions Before
and After Natural Log Transformation of the Data
Before transformation After transformation
Shapiro-Wilk P Shapiro-Wilk P
Muscle Side test statistic value test statistic value
FD Left 0.804 ⬍0.001 0.930 0.048
QD Left 0.838 ⬍0.001 0.918 0.027
TA Left 0.839 ⬍0.001 0.981 0.869
AH Left 0.884 0.003 0.961 0.326
FD Right 0.826 ⬍0.001 0.938 0.078
QD Right 0.804 ⬍0.001 0.960 0.323
TA Right 0.779 ⬍0.001 0.954 0.226
AH Right 0.836 ⬍0.001 0.939 0.088
Muscle abbreviations: FD indicate first dorsal interosseous; QD, quadriceps;
TA, tibialis anterior; AH, abductor halluces.

Table 5. Results of Shapiro-Wilk Tests for

Normality of Baseline Transcranial Electric Motor-
Evoked Potential Latency Distributions Before and
After Natural Log Transformation of the Data
Before transformation After transformation
Shapiro-Wilk P Shapiro-Wilk P
Muscle Side test statistic value test statistic value
FD Left 0.933 0.060 0.953 0.200
QD Left 0.897 0.009 0.939 0.097
TA Left 0.928 0.048 0.952 0.202
AH Left 0.948 0.145 0.961 0.333 Figure 1. Probability density estimates for baseline transcranial
FD Right 0.945 0.124 0.963 0.363 electric motor-evoked potential (tceMEP) amplitudes recorded from
QD Right 0.864 0.001 0.911 0.018 left tibialis anterior muscle in intrathecal morphine and control
TA Right 0.877 0.003 0.932 0.060 subjects. Upper and lower panels show probability distributions
AH Right 0.947 0.138 0.970 0.551 before and after natural log transformation of the amplitude data,
Muscle abbreviations: FD indicate first dorsal interosseous; QD, quadriceps;
respectively. Each panel shows the scaled relative frequency histo-
TA, tibialis anterior; AH, abductor halluces. gram, an estimated fit to the data using a Gaussian kernel smooth-
ing function (solid line), and the normal distribution defined by the
mean and SD of the tceMEP amplitudes (dashed line).
hypothesis was rejected for 2 of 8 amplitude distributions and
1 of 8 latency distributions. The normalizing effect of log
transformation on tceMEP amplitudes is illustrated graphi- Table 6. Results of Levene’s Test for Equality of
cally in Figure 1. Variance of Intrathecal Morphine (ITM) and Control
Levene’s test was used to evaluate the null hypothesis
Group Transcranial Electric Motor-Evoked Potential
Amplitudes Measured in the Posttreatment Period
that control and ITM group population variances were
Before transformation After transformation
equal for the 8 muscle datasets. The results for amplitude
and latency measures are summarized in Tables 6 and 7, F P F P
respectively. The null hypothesis of equal population vari- Muscle Side ratio df value ratio df value
ances was not rejected (i.e., P ⬎ 0.05) either for raw or log FD Left 7.221 1, 28 0.012 3.839 1, 28 0.060
QD Left 0.481 1, 27 0.494 0.387 1, 27 0.539
transformed amplitude measures in 7 of 8 comparisons. A TA Left 0.206 1, 27 0.654 0.001 1, 27 0.973
similar result was obtained with latency measures. AH Left 2.255 1, 28 0.144 0.088 1, 28 0.769
The primary null hypothesis for analyses of tceMEP FD Right 2.639 1, 28 0.115 6.778 1, 28 0.015
amplitudes was that mean amplitudes would be the same QD Right 0.084 1, 27 0.774 0.811 1, 27 0.376
TA Right 3.250 1, 27 0.083 2.411 1, 27 0.132
for ITM and control subjects (i.e., there would be no
AH Right 0.042 1, 28 0.839 0.628 1, 28 0.435
significant main or interaction effects involving treatment
Muscle abbreviations: FD indicate first dorsal interosseous; QD, quadriceps;
group). There were no significant differences in the baseline TA, tibialis anterior; AH, abductor halluces.
response amplitudes of the 2 groups for any of the 4
muscles studied on the left and right sides. ANOVA of
baseline response amplitudes (2 treatment groups ⫻ 4 the 2 treatment groups are shown in Figure 2. Geometric
muscle types ⫻ 2 sides) showed no significant main effect mean differences in the baseline response amplitudes of ITM
of treatment group or side of response, and no significant and control group subjects and associated 95% CIs for each
interaction effects. As expected, there was a main effect of muscle are summarized in Table 8. Average response ampli-
muscle type (P ⫽ 0.033), consistent with commonly ob- tudes collapsed across all muscles for each subject were not
served differences in the amplitudes of tceMEPs recorded significantly different between the two groups during the
from different myotomes. Baseline response amplitudes for baseline period (95% CI) ⫽ ⫺38% to 45%; P ⫽ 0.783).

July 2012 • Volume 115 • Number 1 www.anesthesia-analgesia.org 163

Intrathecal Morphine Effects on Motor-Evoked Potentials

Table 7. Results of Levene’s Test for Equality of Table 8. Percent Difference Between Baseline
Variance of Intrathecal Morphine and Control Transcranial Electric Motor-Evoked Potential
Group Transcranial Electric Motor-Evoked Potential Geometric Mean Amplitudes of Intrathecal
Latencies Measured in the Posttreatment Period Morphine (ITM) And Control Group Subjects, and
Before transformation After transformation Associated 95% Confidence Intervals for
Responses Recorded From Each of 8 Muscles
F P F P
Muscle Side ratio df value ratio df value Percent difference Calculated
Muscle Side (95% CI) P value
FD Left 0.172 1, 28 0.682 0.260 1, 28 0.614
QD Left 2.217 1, 27 0.148 1.215 1, 27 0.280 FD Left 8 (⫺76 to 395) 0.876
TA Left 1.013 1, 27 0.323 0.708 1, 27 0.407 QD Left 53 (⫺45 to 325 0.232
AH Left 0.677 1, 28 0.418 0.439 1, 28 0.513 TA Left ⫺36 (⫺72 to 46) 0.122
FD Right 8.091 1, 28 0.008 6.269 1, 28 0.018 AH Left ⫺14 (⫺68 to 134) 0.665
QD Right 0.974 1, 27 0.332 2.451 1, 27 0.129 FD Right ⫺37 (⫺85 to 137) 0.299
TA Right 1.765 1, 27 0.195 2.384 1, 27 0.134 QD Right 31 (⫺67 to 421) 0.571
AH Right 1.765 1, 28 0.195 1.784 1, 28 0.192 TA Right ⫺16 (⫺62 to 85) 0.504
AH Right 4 (⫺65 to 206) 0.912
Muscle abbreviations: FD indicate first dorsal interosseous; QD, quadriceps;
TA, tibialis anterior; AH, abductor halluces. Confidence intervals are corrected for the effects of multiple comparisons.
Bonferroni correction was applied to define an overall Type I error of 0.05 and
an appropriate level of significance was defined as P ⬍ 0.00625 for each
t-test comparison. Positive and negative percent differences respectively
reflect percent increases and decreases of ITM group means relative to
controls. Muscle abbreviations: FD indicate first dorsal interosseous; QD,
quadriceps; TA, tibialis anterior; AH, abductor halluces.

Figure 2. Baseline transcranial electric motor-evoked potential am-

plitudes for subjects in the intrathecal morphine (n ⫽ 14) and control
(n ⫽ 16) groups. Upper and lower panels summarize results for
responses from muscles on the left and right sides of the body,
respectively. Error bars show 95% confidence interval about the
geometric mean for each group. Muscle abbreviations: FD indicate
first dorsal interosseous; QD, quadriceps; TA, tibialis anterior; AH,
abductor halluces.
As was the case for baseline responses, mean response
amplitudes over the 30-minute period after administration
of ITM did not differ significantly from controls over a
similar time period. Once again, ANOVA revealed a single
main effect of muscle type (P ⬍ 0.001). Mean posttreatment
tceMEP amplitudes for ITM and control group subjects are
shown in Figure 3, and the results of group comparisons for
each of 8 muscles are summarized in Table 9. In all cases,
the lower end of the 95% CI for the mean amplitude
Figure 3. Posttreatment transcranial electric motor-evoked potential
(tceMEP) amplitudes for subjects in the intrathecal morphine and
control groups. Muscle abbreviations: FD indicate first dorsal in-
terosseous; QD, quadriceps; TA, tibialis anterior; AH, abductor
halluces.
difference was short of the primary endpoint of 70%
attenuation specified both in the hypothesis for clinically
significant change and in the statistical power analysis for
the study. For the remaining 2 comparisons, the lower end
of the 95% CI was 70% and 75% attenuation, respectively.
Average response amplitudes collapsed across all muscles
for each subject were not significantly different between the

164 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

 Table 9. Percent Difference Between
Posttreatment Transcranial Electric Motor-Evoked
Potential Geometric Mean Amplitudes of Intrathecal
Morphine (ITM) and Control Group Subjects
Percent difference Calculated
Muscle Side (95% CI) P value
FD Left 16 (⫺65 to 282) 0.719
QD Left 57 (⫺52 to 408) 0.265
TA Left ⫺20 (⫺69 to 105) 0.497
AH Left 35 (⫺48 to 248) 0.360
FD Right ⫺22 (⫺75 to 139) 0.503
QD Right 41 (⫺70 to 552) 0.509
TA Right 4 (⫺57 to 152) 0.898
AH Right 34 (⫺50 to 256) 0.390
Confidence intervals are corrected for the effects of multiple comparisons.
Bonferroni correction was applied to define an overall Type I error of 0.05 and
an appropriate level of significance was defined as P ⬍ 0.00625 for each
t-test comparison. Positive and negative percent differences respectively
reflect percent increases and decreases of ITM group means relative to
controls. Muscle abbreviations: FD indicate first dorsal interosseous; QD,
quadriceps; TA, tibialis anterior; AH, abductor halluces.

ITM group and the control group during the posttreatment

period (95% CI ⫽ ⫺30% to 78%; P ⫽ 0.640).
To address possible effects of chance baseline differ-
ences between the ITM and control groups, mean response
amplitudes in the posttreatment period were reanalyzed
using baseline amplitude as a covariate. ANCOVA con-
firmed absence of any main or interaction effects involving Figure 4. Percent change from baseline in posttreatment transcra-
treatment group. nial electric motor-evoked potential (tceMEP) amplitudes. Asterisks
A third analysis observed effects of ITM on relative indicate statistically significant differences between intrathecal mor-
changes in tceMEP amplitude from baseline. ⬎96% of the phine and control groups (P ⫽ 0.039 for left AH response, P ⫽ 0.048
for right AH response). Muscle abbreviations: FD indicate first dorsal
recorded baseline amplitudes in this study exceeded 50 interosseous; QD, quadriceps; TA, tibialis anterior; AH, abductor
micro volts, which was used as the minimum criterion for halluces.
inclusion of these data in analyses of relative amplitude.
The mean relative tceMEP amplitudes for ITM and control
group subjects are shown in Figure 4. ANOVA revealed Table 10. Percent Difference Between Transcranial
both a significant main effect of muscle type (P ⫽ 0.028) and Electric Motor-Evoked Potential Geometric Mean
a significant interaction between treatment group and Amplitude Ratios (i.e., Posttreatment/Baseline) of
muscle type (P ⫽ 0.050). The significant interaction effect Intrathecal (ITM) and Control Group Subjects
was explored with post hoc t-test comparisons of ITM and Percent difference Calculated
Muscle Side (95% CI) P value
control responses for each of 8 muscles, as summarized in
FD Left 46 (⫺37 to 239) 0.198
Table 10. The null hypothesis of no significant difference
QD Left ⫺3 (⫺22 to 22) 0.735
between ITM and control response amplitudes was not TA Left 26 (⫺33 to 137) 0.296
rejected for any of the 8 comparisons, using a Bonferroni- AH Left 57 (⫺16 to 194) 0.039
adjusted cutoff P value of 0.00,625 for each comparison. The FD Right 32 (⫺18 to 113) 0.098
smallest calculated P values were for t-test comparisons of QD Right 1 (⫺31 to 48) 0.913
TA Right 25 (⫺24 to 106) 0.188
ITM and control responses from left and right abductor AH Right 57 (⫺18 to 201) 0.048
hallucis muscles (P ⫽ 0.039 and P ⫽ 0.048, respectively). On
Confidence intervals are corrected for the effects of multiple comparisons.
average, ITM response amplitudes recorded from left ab- Bonferroni correction was applied to define an overall Type I error of 0.05 and
ductor hallucis muscle exceeded baseline by 49% (95% CI ⫽ an appropriate level of significance was defined as P ⬍ 0.00625 for each
7% to 107%) compared to a 6% (CI ⫽ ⫺26% to 20%) t-test comparison. Positive and negative percent differences respectively
reflect percent increases and decreases of ITM group means relative to
decrease noted for the corresponding muscle in the control controls. Muscle abbreviations: FD indicate first dorsal interosseous; QD,
group (P ⫽ 0.039). Similarly, ITM group response ampli- quadriceps; TA, tibialis anterior; AH, abductor halluces.
tudes for the right abductor hallucis muscle exceeded
baseline by 32% (CI ⫽ 4% to 66%) compared to a 16% (CI ⫽ in Figure 4. Ninety-five percent confidence intervals about
⫺41% to 20%) decrease for the control group (P ⫽ 0.048). the mean included 0% change from baseline for all but the
Abductor hallucis muscle tceMEPs showed an increase left and right abductor hallucis muscle responses of the
from baseline as soon as 5 minutes after ITM administra- ITM group.
tion, and peaked for both muscles at 20 min, as shown in Analysis of tceMEP onset latencies revealed no signifi-
Figure 5. Response amplitudes recorded from the remain- cant differences between ITM and control subjects either in
ing muscles in the posttreatment period did not differ from the baseline or posttreatment periods. Mean group laten-
baseline, either for ITM or control subjects, as summarized cies in the baseline and posttreatment periods are shown in

July 2012 • Volume 115 • Number 1 www.anesthesia-analgesia.org 165

Intrathecal Morphine Effects on Motor-Evoked Potentials

Figure 5. Percent change from baseline in left and right abductor

hallucis muscle transcranial electric motor-evoked potential (tce-
MEP) amplitudes at 5, 10, 20, and 30 minutes after administration
of intrathecal morphine (ITM). Error bars show 95% confidence
intervals about geometric means.

Figure 7. Posttreatment transcranial electric motor-evoked potential

(tceMEP) latencies for subjects in the intrathecal morphine and
control groups. Muscle abbreviations: FD indicate first dorsal in-
terosseous; QD, quadriceps; TA, tibialis anterior; AH, abductor
halluces.

Table 11. Percent Difference Between Baseline

Transcranial Electric Motor-Evoked Potential
Geometric Mean Latencies of Intrathecal Morphine
(ITM) and Control Group Subjects.
Percent difference Calculated
Muscle Side (95% CI) P value
FD Left 2 (⫺14 to 20) 0.760
QD Left 5 (⫺11 to 25) 0.380
TA Left 4 (⫺7 to 17) 0.270
AH Left 6 (⫺5 to 17) 0.147
FD Right 4 (⫺10 to 20) 0.450
QD Right 6 (⫺14 to 29) 0.425
TA Right 7 (⫺7 to 23) 0.155
AH Right 3 (⫺7 to 15) 0.346
Confidence intervals are corrected for the effects of multiple comparisons.
Bonferroni correction was applied to define an overall Type-I error of 0.05 and
Figure 6. Baseline transcranial electric motor-evoked potential (tce- an appropriate level of significance was defined as P ⬍ 0.00625 for each
MEP) latencies for subjects in the intrathecal morphine and control t-test comparison. Positive and negative percent differences respectively
groups. Muscle abbreviations: FD indicate first dorsal interosseous; reflect percent increases and decreases of ITM group means relative to
QD, quadriceps; TA, tibialis anterior; AH, abductor halluces. controls. Muscle abbreviations: FD indicate first dorsal interosseous; QD,
quadriceps; TA, tibialis anterior; AH, abductor halluces.

Figures 6 and 7, respectively. ANOVA showed no main
effects of treatment group and no significant treatment
group interactions. The only significant main effect was for
muscle type (P ⬍ 0.001 for both baseline and mean post-
treatment latencies), as expected based on length of neural
pathways to different muscles. Statistical comparisons of
ITM and control group latencies in the baseline and post-
treatment periods are summarized for each of 8 muscles in
Tables 11 and 12, respectively. Average response latencies
collapsed across all muscles for each subject were 4% larger
for the ITM group than for controls during the baseline
period (95% CI ⫽ ⫺5% to 13%; P ⫽ 0.377), and 3% larger for
the ITM group than for controls during the posttreatment
period (95% CI ⫽ ⫺4% to 12%; P ⫽ 0.359).
ANCOVA using baseline latency as the covariate con-
firmed a lack of significant treatment effects. Baseline and
posttreatment response latencies were generally compa-
rable for any given muscle, as summarized in Figure 8.

166 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

 Table 12. Percent Difference Between Table 13. Percent Difference Between
Posttreatment Transcranial Electric Motor-Evoked Transcranial Electric Motor-Evoked Potential
Potential Geometric Mean Latencies of Intrathecal Geometric Mean Latency Ratios (i.e.
Morphine (ITM) and Control Group Subjects Posttreatment/Baseline) of Intrathecal Morphine
Percent difference Calculated (ITM) and Control Group Subjects
Muscle Side (95% CI) P value Percent difference Calculated
FD Left 2 (⫺13 to 20) 0.684 Muscle Side (95% CI) P value
QD Left 5 (⫺9 to 22) 0.289 FD Left 1 (⫺4 to 6) 0.599
TA Left 4 (⫺6 to 15) 0.294 QD Left 1 (⫺8 to 10) 0.865
AH Left 3 (⫺6 to 14) 0.336 TA Left ⫺1 (⫺4 to 3) 0.541
FD Right 7 (⫺7 to 24) 0.155 AH Left ⫺2 (⫺6 to 1) 0.043
QD Right 6 (⫺10 to 25) 0.329 FD Right 3 (⫺1 to 8) 0.053
TA Right 6 (⫺6 to 19) 0.166 QD Right 0 (⫺1 to 8) 0.993
AH Right 4 (⫺8 to 16) 0.378 TA Right ⫺1 (⫺5 to 3) 0.346
Confidence intervals are corrected for the effects of multiple comparisons. AH Right 0 (⫺1 to 2) 0.572
Bonferroni correction was applied to define an overall Type I error of 0.05 and Confidence intervals are corrected for the effects of multiple comparisons.
an appropriate level of significance was defined as P ⬍ 0.00625 for each Bonferroni correction was applied to define an overall Type I error of 0.05 and
t-test comparison. Positive and negative percent differences respectively an appropriate level of significance was defined as P ⬍ 0.00625 for each
reflect percent increases and decreases of ITM group means relative to t-test comparison. Positive and negative percent differences respectively
controls. Muscle abbreviations: FD indicate first dorsal interosseous; QD, reflect percent increases and decreases of ITM group means relative to
quadriceps; TA, tibialis anterior; AH, abductor halluces. controls. Muscle abbreviations: FD indicate first dorsal interosseous; QD,
quadriceps; TA, tibialis anterior; AH, abductor halluces.

subjects, with the possible exception of responses recorded

Figure 8. Percent change from baseline in posttreatment transcra-
nial electric motor-evoked potential (tceMEP) latencies. Muscle
abbreviations: FD indicate first dorsal interosseous; QD, quadriceps;
TA, tibialis anterior; AH, abductor halluces.
ANOVA for relative response latency showed no signifi-
cant main or interaction effects for treatment group, muscle
type or side of response. Statistical comparisons of ITM and
control group relative latencies are summarized for each of
8 muscles in Table 13.
DISCUSSION
The results of this study show that mean tceMEP ampli-
tudes and latencies in the 30-minute period after ITM
injection were comparable to those recorded for control
from left and right abductor hallucis muscles, which appeared
to increase in amplitude after administration of ITM. The
relatively few statistically significant differences noted in this
study did not rise to the level of clinical significance. Further
study is suggested to determine whether ITM may induce
changes in tceMEPs after the initial 30 minutes after injection.
This information would give confidence to anesthesiolo-
gists injecting ITM before the start of surgery that this
technique will not interfere with intraoperative neuro-
physiologic monitoring. Earlier injection of ITM allows full
onset of analgesic effects before emergence; it may also
provide preemptive analgesic effects, although this is un-
proven. In addition, use of ITM in PSF has been shown to
decrease the amount of blood loss during surgery, an effect
that appears to be mediated through arterial blood
pressure-decreasing effects.14 The potential impact of de-
creased blood pressure on neurophysiologic monitoring
should also be considered.
In the present study, motor-evoked potentials were
recorded bilaterally from 4 muscles commonly used by our
group to monitor corticospinal tract conduction during PSF
surgery for idiopathic scoliosis. Responses recorded from
quadriceps, tibialis anterior, and abductor hallucis muscles
provide monitoring coverage for the spinal cord and L3-S1
spinal nerve roots, whereas those from first dorsal in-
terosseous muscle assay the cervical spinal cord, brachial
plexus, and other upper extremity peripheral nerve fibers,
in addition to serving as controls. tceMEPs from the first
dorsal interosseous muscle would have been least prone to
any potential effects of ITM, based on distance of their
innervating C8-T1 roots from the site of ITM injection.
Transient paraparesis in a woman undergoing thoracoab-
dominal aneurysm repair after neuraxial morphine injection
that was reversed by administration of naloxone has been
described.10 This phenomenon was reproduced in a rat model
after transient spinal cord ischemia.9 –10 In a rat spinal cord
contusion model, ITM impaired recovery and increased the
size of the injury lesion.11 The synthetic opioids fentanyl,

July 2012 • Volume 115 • Number 1 www.anesthesia-analgesia.org 167

Intrathecal Morphine Effects on Motor-Evoked Potentials
alfentanil, and sufentanil have been shown to cause a dose-
dependent suppression of motor-evoked potentials after IV
administration in rabbits.15 Given these findings, this study
was conducted as an initial investigation into the effects of
ITM on tceMEPs in human subjects. Although our results are
encouraging, it is yet unknown whether the same results
would be observed if ITM were injected after an operation
where there was a transient episode of spinal cord ischemia or
contusion.
Limitations
This study has a number of limitations. Importantly, the
study addressed tceMEP changes in the first 30 minutes
after ITM injection, which precludes generalization beyond
that time period. The study had sufficient statistical power
to identify a clinically significant effect size of 70% ampli-
tude attenuation after administration of ITM; attenuation
⬍70%, if present, may have gone undetected. This is not a
randomized or placebo-controlled trial. Blinding during
data interpretation mitigates this limitation. Another poten-
tial confounder is that study group assignment was deter-
mined by the caudal extent of the operation, which creates
a difference in the surgical procedures performed in the 2
groups, although there is no plausible reason to believe that
this variable could have affected the study outcome. In an
effort to limit any effect from this potentially confounding
variable, only subjects who had uneventful surgical correc-
tions without any clinically significant intraoperative tce-
MEP signal changes were included in this investigation.
Moreover, the overall number of vertebral levels operated
on was the same for experimental and control groups.
Overall, these limitations do not detract from the primary
findings of the study.
CONCLUSIONS
ITM administered in the doses used in this investigation
did not cause a ⬎70% attenuation of mean tceMEP
amplitudes or latency changes during the 30-minute
period immediately after injection. Further studies are
required to determine if there are any delayed effects
after this initial time period.
DISCLOSURES
Name: Paul A. Stricker, MD.
Contribution: This author helped design the study, conduct
the study, collect the data, and prepare the manuscript.
Conflicts of Interest: Paul A. Stricker has no conflict of
interest.
Name: Anthony K. Sestokas, PhD, D. ABNM, F. ASNM.
Contribution: This author helped design the study, conduct
the study, collect the data, and prepare the manuscript.
Conflicts of Interest: Anthony K. Sestokas has no conflict of
interest.
Name: Daniel Schwartz, PhD, D. ABNM, F. ASNM.
Contribution: This author helped conduct the study, collect
the data, analyze the data, and prepare the manuscript.
Conflicts of Interest: Daniel Schwartz has no conflict of
interest.
Name: Vidya Bhalodia, MS, D. ABNM.
Contribution: This author helped conduct the study, collect
the data, analyze the data, and prepare the manuscript.
Conflicts of Interest: Vidya Bhalodia has no conflict of interest.
168 www.anesthesia-analgesia.org
Name: Ankit Pahwa, MS.
Contribution: This author helped analyze the data and pre-
pare the manuscript.
Conflicts of Interest: Ankit Pahwa has no conflict of interest.
Name: John P. Dormans, MD, FACS.
Contribution: This author helped design the study and con-
duct the study.
Conflicts of Interest: John P. Dormans has no conflict of
interest.
Name: Lia L. LaBrant, BA.
Contribution: This author helped conduct the study and
collect the data.
Conflicts of Interest: Lia L. LaBrant has no conflict of interest.
Name: Beverly J. Burgess, BA.
Contribution: This author helped conduct the study and
collect the data.
Conflicts of Interest: Beverly J. Burgess has no conflict of
interest.
Name: F. Wickham Kraemer, MD.
Contribution: This author helped design the study, conduct
the study, and prepare the manuscript.
Conflicts of Interest: F. Wickham Kraemer has no conflict of
interest.
Name: Arjunan Ganesh, MBBS.
Contribution: This author helped design the study, conduct
the study, and prepare the manuscript.
Conflicts of Interest: Arjunan Ganesh has no conflict of
interest.
This manuscript was handled by: Gregory Crosby, MD.
REFERENCES
1. Kawaguchi M, Sakamoto T, Inoue S, Kakimoto M, Furuya H,
Morimoto T, Sakaki T. Low dose propofol as a supplement to
ketamine-based anesthesia during intraoperative monitoring
of motor-evoked potentials. Spine (Phila Pa) 2000;25:974 –9
2. Calancie B, Klose KJ, Baier S, Green BA. Isoflurane-induced
attenuation of motor evoked potentials caused by electrical
motor cortex stimulation during surgery. J Neurosurg
1991;74:897–904
3. Haghighi SS, Sirintrapun SJ, Keller BP, Oro JJ, Madsen R. Effect
of desflurane anesthesia on transcortical motor evoked poten-
tials. J Neurosurg Anesthesiol 1996;8:47–51
4. Haghighi SS, Green KD, Oro JJ, Drake RK, Kracke GR. Depres-
sive effect of isoflurane anesthesia on motor evoked potentials.
Neurosurgery 1990;26:993–7
5. Jellinek D, Platt M, Jewkes D, Symon L. Effects of nitrous oxide
on motor evoked potentials recorded from skeletal muscle in
patients under total anesthesia with intravenously adminis-
tered propofol. Neurosurgery 1991;29:558 – 62
6. Kalkman CJ, Drummond JC, Ribberink AA, Patel PM, Sano
T, Bickford RG. Effects of propofol, etomidate, midazolam,
and fentanyl on motor evoked responses to transcranial
electrical or magnetic stimulation in humans. Anesthesiol-
ogy 1992;76:502–9
7. Kalkman CJ, Drummond JC, Patel PM, Sano T, Chesnut RM.
Effects of droperidol, pentobarbital, and ketamine on myo-
genic transcranial magnetic motor-evoked responses in hu-
mans. Neurosurgery 1994;35:1066 –71
8. Goodarzi M, Shier NH, Grogan DP. Effect of intrathecal
opioids on somatosensory-evoked potentials during spinal
fusion in children. Spine (Phila Pa) 1996;21:1565– 8
9. Kakinohana M, Fuchigami T, Nakamura S, Sasara T, Kawa-
bata T, Sugahara K. Intrathecal administration of morphine,
but not small dose, induced spastic paraparesis after a
noninjurious interval of aortic occlusion in rats. Anesth
Analg 2003;96:769 –75
ANESTHESIA & ANALGESIA
10. Kakinohana M, Marsala M, Carter C, Davison JK, Yaksh TL.
Neuraxial morphine may trigger transient motor dysfunction
after a noninjurious interval of spinal cord ischemia: a clinical
and experimental study. Anesthesiology 2003;98:862–70
11. Hook MA, Moreno G, Woller S, Puga D, Hoy K, Jr., Balden R,
Grau JW. Intrathecal morphine attenuates recovery of function
after a spinal cord injury. J Neurotrauma 2009;26:741–52
12. Schwartz DM, Auerbach JD, Dormans JP, Flynn J, Drummond
DS, Bowe JA, Laufer S, Shah SA, Bowen JR, Pizzutillo PD,
Jones KJ. Neurophysiological detection of impending spinal
cord injury during scoliosis surgery. J Bone Joint Surg Am
2007;89:2440 –9
13. Mahmoud M, Sadhasivam S, Salisbury S, Nick TG, Schnell B,
Sestokas AK, Wiggins C, Samuels P, Kabalin T, McAuliffe J.
Susceptibility of transcranial electric motor-evoked potentials
to varying targeted blood levels of dexmedetomidine during
spine surgery. Anesthesiology 2010;112:1364 –73
14. Gall O, Aubineau JV, Berniere J, Desjeux L, Murat I. Analgesic
effect of low-dose intrathecal morphine after spinal fusion in
children. Anesthesiology 2001;94:447–52
15. Thees C, Scheufler KM, Nadstawek J, Pechstein U, Hanisch M,
Juntke R, Zentner J, Hoeft A. Influence of fentanyl, alfentanil,
and sufentanil on motor evoked potentials. J Neurosurg Anes-
thesiol 1999;11:112– 8

Protocolized Intensive Care Unit Management of Analgesia, Sedation, and Delirium Improves
Analgesia and Subsyndromal Delirium Rates: Erratum

In the article that appeared on page 451 in the August 2010 issue of volume 111 of Anesthesia & Analgesia,
the authors would like to correct an error in the Appendix. The bedside tool (Appendix) erroneously
states that delirium assessment was performed with RASS levels between ⫺2 and ⫹4, which reflects our
usual clinical care. However, in the study, all patients with RASS levels between ⫺3 and ⫹4 were
evaluated for delirium.
Reference:
Skrobik Y, Ahern S, Leblanc M, Marquis F, Awissi DK, Kavanagh BP. Protocolized intensive care unit
management of analgesia, sedation, and delirium improves analgesia and subsyndromal delirium rates.
Anesth Analg 2010;111:451– 63

July 2012 • Volume 115 • Number 1 www.anesthesia-analgesia.org 169