Sensory and Affective Pain Discrimination After Inhalation of Essential Oils

JEFFREY J. GEDNEY, PSYD, TONI L. GLOVER, MA, RN, AND ROGER B. FILLINGIM, PHD
Objective: The purpose of this investigation was to examine the effects of olfactory absorption of two commonly used therapeutic
essential oils on sensory and affective responses to experimentally induced pain. Methods: A sex-balanced (13 men and 13 women)
randomized crossover design was used to obtain pre- and posttreatment change scores for quantitative sensory ratings of contact
heat, pressure, and ischemic pain across separate inhalation treatment conditions using essential oil of lavender, essential oil of
rosemary, and distilled water (control). Subjective reports of treatment-related changes in pain intensity and pain unpleasantness
were obtained for each condition using a visual analog scale. We interpret our findings with respect to the separate dimensions of
sensory and affective processing of pain. Results: Analyses revealed the absence of changes in quantitative pain sensitivity ratings
between conditions. However, retrospectively, subjects’ global impression of treatment outcome indicated that both pain intensity
and pain unpleasantness were reduced after treatment with lavender and marginally reduced after treatment with rosemary,
compared with the control condition. Conclusion: These findings suggest that aromatherapy may not elicit a direct analgesic effect
but instead may alter affective appraisal of the experience and consequent retrospective evaluation of treatment-related pain. Key
words: aromatherapy, sensory testing, pain intensity, pain unpleasantness.

PANAS ⫽ Positive Affect and Negative Affect Scale; RIA ⫽ barium enema suspension, compared with an equal number of
radioimmunoassay; STAI ⫽ State Trait Anxiety Index; VAS ⫽ patients (n ⫽ 70) receiving barium suspension without pep-
visual analog scales. permint oil (10). Woolfson and Hewitt randomized 36 inten-
sive care unit patients to one of three groups: massage with
INTRODUCTION
lavender, massage without lavender, or rest only (21). Treat-

T he use of complementary therapies as adjuncts to tradi-

tional Western medical practice has significantly in-
creased in the United States over the past 2 decades. Rising
health care costs and the increasing acceptance of holistic
approaches to health care have led to acceptance and use of
previously marginalized therapeutic modalities by both the
general patient population and healthcare professionals. Aro-
matherapy is the fastest growing complementary therapy (1).
Aromatherapy includes the inhaled, absorbed, or ingested use
of essential oil extracts from plants and flowers for prophy-
lactic medical care or active treatment. A growing body of
clinical and laboratory studies report beneficial effects of
essential oils on physiological and psychological processes in
both animals and humans (2–20). Presently, two experimental
animal studies demonstrate physiological effects of essential
oils. Buchbauer and colleagues showed that under standard-
ized laboratory conditions, male and female mice displayed a
significant decrease in activity on inhalation of oil of lavender
(20). In a second experiment, a hyperactive state was induced
by injection of caffeine, after which inhalation of oil of lav-
ender produced a sedative effect, significantly reducing the
mice’s motility to nearly their baseline level. In the second
published study, Lis-Balchin and colleagues topically applied
oil of lavender to an in vitro preparation of guinea-pig ileum
after electrical stimulation induced spasm and observed a
significant decrease in spasmodic action (19).
Three well-designed clinical studies have been published
supporting the benefits of aromatherapy for general medical
management. Sparks and colleagues demonstrated that pep-
permint oil significantly reduced colon spasm when added to
From the Division of Public Health Services and Research, University of
Florida College of Dentistry, Gainesville, FL.
Address correspondence and reprint requests to Dr. Roger Fillingim, 1600
SW Archer Road, Box 100404, Gainesville, FL 32610. E-mail:
rfillingim@dental.ufl.edu
Received for publication October 23, 2003; revision received March 26,
2004.
DOI: 10.1097/01.psy.0000132875.01986.47
ment consisted of 20 minutes of foot massage twice a week for
5 weeks. A significantly greater number of patients in the
lavender massage group demonstrated decreased heart rate,
compared with the other two groups. These results suggest
that oil of lavender reduces physiological arousal above that
provided by massage alone or rest only. Similar conclusions
were obtained by Buckle, who recorded emotional and behav-
ioral stress levels during a randomized double-blind trial using
two different species of lavender massaged for 20 minutes on
2 consecutive days into the extremities and forehead of two
groups of post– cardiac bypass patients (N ⫽ 28) (22). A
significant difference in self-reported stress between the two
groups was attributed to the differential therapeutic effect
between the two species of lavender oil, and not to the effect
of massage because this was common to all treatment condi-
tions.
The relaxant effect of oil of lavender has also been dem-
onstrated in experimental studies with humans. Diego and
colleagues used standardized measures of electroencephalo-
gram (EEG) activity, alertness, and mood for 40 adults after
inhalation of either oil of lavender or oil of rosemary (con-
sidered a stimulating odor) (9). Subjects were given simple
math computations before and after the therapy. Subjects who
received the lavender aroma reported a significantly improved
mood, displayed a significant improvement on math compu-
tation accuracy, and exhibited a significant increase in EEG
frontal alpha power, suggesting increased drowsiness. In con-
trast, the rosemary aroma group showed a decrease in EEG
frontal alpha power and beta power, suggesting increased
alertness. Moss and colleagues also used essential oils of
lavender and rosemary in a study of cognitive performance
and mood in which 144 volunteers were randomly assigned to
one of three groups: lavender, rosemary, or control (23).
Subjects were deceived as to the genuine aim of the study and
were asked to complete a computerized cognitive assessment
battery. Visual analog mood questionnaires were completed
before exposure to each condition and after completion of the

Psychosomatic Medicine 66:599 – 606 (2004) 599

0033-3174/04/6604-0599
Copyright © 2004 by the American Psychosomatic Society

cognitive assessment battery. They found that lavender pro-
duced a decrease in performance of attention and working
memory capacity and reaction times compared with controls.
Rosemary enhanced memory performance but reduced mem-
ory reaction times compared with controls. In terms of posttest
mood ratings, the control group was significantly less content
than both the rosemary and lavender groups. Other studies
have similarly reported improved memory performance (2,3)
and alertness (7) after olfactory exposure to different odors.
Very few studies have investigated the effects of aroma-
therapy in the treatment and management of pain (5,6,8,24).
Gobel and colleagues used a crossover design to compare
treatment effect between inhaled peppermint, acetamino-
phen, peppermint plus acetaminophen, and placebo for
headache pain in 41 adult males (11). Peppermint and
acetaminophen were equally effective in reducing headache
pain and both were significantly better than placebo. Gobel
et al. topically administered preparations of peppermint oil,
eucalyptus, and ethanol to the forehead and temples of 32
healthy males, using a double-blind, randomized, crossover
design (12). Pre- and postmanipulation measures included
temporal electromyogram (EMG) activity, mood state, and
sensitivity to experimental pain. Significant results in-
cluded decreased emotional irritation, decreased resting
EMG, and diminished ischemic pain only for those prepa-
rations containing peppermint oil. Marchand and Arsenault
investigated the effects of pleasant, unpleasant, and neutral
odors on mood and pain ratings (6). Twenty men and 20
women individually rank ordered a series of common odors
(eg, orange water, vinegar, baby oil) as pleasant, unpleas-
ant, or neutral. They then completed a series of 3-minute
hand immersions in a hot circulating bath (46 – 48°C) while
continually smelling one of the three previously selected
odors (most pleasant, most unpleasant, and closest to neu-
tral). Pain intensity and pain unpleasantness ratings were
reported every 15 seconds during each trial. Women per-
ceived significantly less pain intensity and pain unpleas-
antness during inhalation of the pleasant odor relative to the
unpleasant or neutral odors. Pain intensity and unpleasant-
ness did not differ across odor conditions for men. These
findings suggest a sex-dependent effect of pleasant odor on
perceived pain intensity and unpleasantness. It remains to
be seen whether therapeutic essential oils produce similar
effects.
The purpose of this investigation was to examine the ef-
fects of olfactory absorption of two commonly used therapeu-
tic essential oils on sensory and affective responses to exper-
imentally induced pain. We hypothesized that inhalation of
essential oil of lavender and rosemary, both essential oils with
purported analgesic properties (1,8) but opposite effects on
affective arousal, would decrease pain in response to noxious
heat, pressure, and ischemic stimuli. Second, we hypothesized
subjective reports of pain intensity and pain unpleasantness to
be less after separate inhalation treatment sessions with both
essential oil of lavender and rosemary.
600
J. J. GEDNEY et al.
METHODS
A randomized crossover design was used to test for changes in objective
and subjective reports of pain sensitivity after inhalation of commercially
available essential oil of lavender (Lavandula angustifolia) and essential oil of
rosemary (Rosmarinus officinalis) (Elizabeth VanBuren Aromatherapy, Santa
Cruz, CA), species most commonly used in clinical and research applications.
Purity of the oils was documented by the distributor using mass spectrometry
and gas chromatography. Distilled water was used as a control. The three
sessions were separated by 48 to 96 hours and were identical with the
exception of the treatment condition and completion of additional question-
naires during the first session. Quantitative pain measures included contact
heat (threshold, tolerance, and temporal summation of heat pain), pressure
pain threshold (trapezius and masseter), and ischemic pain threshold and
tolerance (circulatory occlusion of the upper arm) pain tasks. Procedures were
identical to those previously reported in the published pain literature (25).
Subjective measures consisted of subjects’ self-reported change in sensory
intensity and pain unpleasantness for each condition.
A male and female experimenter were present during each session (6,26 –
29). To control for hormonal fluctuation believed to influence pain sensitivity
(30), all sessions were conducted during the follicular phase (between days 4
and 10 after menses) (31). To minimize measurement error, the same exper-
imenter obtained measures during each of the three sessions, with male and
female experimenters randomized between subjects.
Subjects
Twenty-six healthy, nonsmoking and unmedicated adult subjects (13 men
and 13 premenopausal women) were recruited from a pool of subjects who
completed a similar experimental pain protocol within the past 18 months.
These subjects were thus familiar with the quantitative sensory testing appa-
ratus and methodology, which reduced novelty effects. Subjects were blinded
to the study’s hypotheses and treatment conditions. The study protocol con-
formed to the Institutional Review Board of the University of Florida. Sub-
jects were paid $75 for completion of the study. All subjects completed the
study.
Session Baseline
On arrival, subjects provided informed consent and were comfortably
seated in a reclining chair in a sound-attenuated room. Subjects then com-
pleted psychological measures assessing affectivity (Positive Affect and Neg-
ative Affect Scale; PANAS (32), measures of trait and situational anxiety
(State Trait Anxiety Index; STAI) (33), and present mood, using the Profile
of Mood States (34). Demographics and health history information and a
questionnaire assessing the subject’s familiarity with and attitudes about
complementary and alternative therapy modalities were completed during the
first session.
Experimental Pain Tasks
Thermal Pain Threshold, Tolerance, and Temporal
Summation Procedures
Thermal pain stimuli were delivered using a computer-controlled Medoc
Thermal Sensory Analyzer (TSA-2001, Ramat Yishai, Israel), which is a
peltier-element-based stimulator using a 3 ⫻ 3-cm thermode. The thermode
was placed on the right dorsal forearm. Its position was altered slightly
between trials in order to avoid either sensitization or habituation of cutaneous
receptors. The maximum permitted temperature was 52°C. Subjects could
terminate the heat stimulus at any point by pressing a control device button or
by verbal command.
Heat pain threshold and tolerance were obtained three times each. On
initiation of each trial, the thermal contact increased in temperature at a rate
of 0.5°C/second from a baseline temperature of 32°C. For pain threshold,
subjects were instructed to press a control button “when the heat first becomes
painful, when you first feel pain.” For pain tolerance, subjects were instructed
to press the button “when you no longer feel able to tolerate the pain.”
Temporal summation of thermal pain was assessed using a series of 10
consecutive 0.5-second-long heat pulses, with an interpulse interval of 2.5
seconds. Separate series of 10 trials with target temperatures of 49°C and
Psychosomatic Medicine 66:599 – 606 (2004)
AROMATHERAPY AND PAIN DISCRIMINATION
52°C were delivered to the right ventral forearm. At each pulse peak, subjects
rated the intensity of the stimulus using a verbal rating scale, with anchors of
0 (no pain) and 100 (most intense pain imaginable). A rating of 20 was
defined as “barely painful.”
Pressure Pain Procedure
Pressure pain threshold was measured using a digital, handheld, clinical
grade pressure algometer (Jtech, Heber City, UT). Pressure was applied at the
rate of 0.5 kg/second to the right upper trapezius and the right masseter, using
a 0.5-cm2 probe. Subjects were instructed to respond verbally when “the
pressure first becomes painful, when you first feel pain.” Three trials were
completed for each site.
Ischemic Pain Procedure
Ischemic pain was assessed using the submaximal effort tourniquet pro-
cedure (35). The ischemic pain task consisted of elevating the right arm above
heart level for 30 seconds. Circulation was then occluded using a standard
blood pressure cuff inflated to 240 mm Hg. The arm was then lowered and the
subject completed 20 hand-grip exercises of 2-second duration at 4-second
intervals at 50% of their maximum grip strength. Subjects were instructed to
report when they first felt pain (pain threshold) and when the pain became
intolerable (pain tolerance), and the time required to reach these two end-
points was recorded. Every 30 seconds after the first report of ischemic pain,
subjects rated in alternate order their ischemic pain intensity and unpleasant-
ness using a 0 to 20 combined verbal–numerical box scales (36). The
procedure was terminated when one of the following occurred: a rating of 20
was obtained, on request by the subject, or after 15 minutes.
Treatment Condition
Treatment consisted of the randomized application of 5 drops of essential
oil of lavender, rosemary, or distilled water to a 2 ⫻ 2-inch cotton gauze. The
gauze was attached to the front of the subject’s garment, approximately 12
inches below their nose. Subjects were directed to breathe normally while
they sat quietly in a recliner for 10 minutes. The gauze remained in place
while subjects then began the posttreatment experimental pain procedures.
This procedure reflects a modification of that used by others (8) by increasing
the application dose and duration of exposure (24).
Salivary Cortisol
Saliva samples for determination of cortisol were obtained to coincide
with the end of the resting period, the end of treatment, and the midpoints of
the pre- and posttreatment pain tasks. Samples were collected 25 minutes after
each of these time points, thus coinciding with the peak cortisol concentration
present in the saliva (37). Samples were collected using an absorbent cotton
pellet (Salivette, Sarstedt Inc., Newton, NC) and frozen at ⫺20°C. Before
assay, saliva was extracted from the Salivettes by thawing and then centrif-
ugation (Beckman refrigerated centrifuge, Model J-6B) at 3000 RPM at 4°C
for 20 minutes. Salivary cortisol levels were determined using the DSL-2000
cortisol radioimmunoassay (RIA) kit (Diagnostic Systems Laboratories, Inc.,
Webster, TX) with modifications. The modifications increased the sensitivity
of the cortisol assay and provided a reliable assay system to detect salivary
cortisol in all samples provided. Modifications of the standard cortisol RIA
TABLE 1. Salivary Cortisol Concentration (␮g/dl) Obtained at Baseline and Immediately Posttreatment, Reported for Sex by Treatment
Condition (Standard deviation reported in parentheses)
Control Lavender
Baseline Treatment Baseline Treatment
Men .183 (.091) .109 (.036) .168 (.075) .120 (.046)
Women .170 (.062) .139 (.091) .163 (.063) .122 (.056)
Psychosomatic Medicine 66:599 – 606 (2004)
procedures were as follows: a) All standards and controls provided were
diluted 1:10 with the cortisol assay dilution buffer. This provided a working
standard curve range of (0.05– 6.00 ␮g/dl); b) All standards, controls and
unknowns were assayed in duplicate at 100 ␮l; c) Cortisol (I125) reagent and
cortisol antiserum complex were added in 100-␮l volumes to all appropriate
tubes; and d) All samples were incubated overnight at room temperature. All
other RIA procedures followed the protocol provided in the DSL-2000
cortisol RIA kit. With the above modifications, this provided an assay system
that had a sensitivity of 0.05 ␮g/dl (at 90% binding) and an intra-assay CV
and interassay CV of 3.95% and 4.94%, respectively.
Postsession Questionnaire
After completing each session’s posttreatment experimental pain proce-
dures, subjects completed a postsession evaluation questionnaire. Using
100-mm visual analog scales (VAS), subjects were asked to rate whether the
session’s odor was relaxing (0-mm anchor) or stimulating (100-mm anchor).
Ratings of 50 mm indicated a neutral rating for each VAS. Also, subjects rated
whether the session’s aroma increased pain intensity and pain unpleasantness
(0-mm anchor) or decreased pain intensity and pain unpleasantness (100-mm
anchor). Finally, subjects rated the intensity of the odor using a 100-mm VAS
with anchors of “not noticeable” (0 mm) and “intense” (100 mm).
Analysis
Between-session differences for baseline anxiety, affectivity, and post-
treatment evaluation questionnaire data were tested using a series of one-way
analysis of variance (ANOVA) and Wilcoxon Z for related, nonparametric
samples. A series of 2 Time (pretreatment, posttreatment) ⫻ 3 Condition
(control, lavender, rosemary) repeated-measures ANOVA were used to test
for time-by-condition interaction effects on sensory test scores, cardiovascular
measures (systolic blood pressure, diastolic blood pressure, heart rate) and
salivary cortisol. Main effects for sex were tested by entering sex as a
between-subjects factor for each ANOVA. Bonferroni adjustments were used
to correct for post hoc pairwise comparisons. Two measures were derived
from the temporal summation procedure. First, the average of all 10 pain
ratings were computed. Then, the maximum increase was calculated as the
difference between the ratings of first heat pulse and the maximum rating
from the remaining pulses. Average ratings of ischemic pain intensity and
unpleasantness were computed by calculating the mean of all ratings for each
subject. For subjects who terminated the task before the time limit, their last
rating was carried forward for all future time points.
RESULTS
Baseline
No differences in baseline measures of salivary cortisol, STAI, or
PANAS emerged between the three treatment conditions (p val-
ues ⬎ .1), demonstrating the absence of a session order effect.
Accordingly, order was not controlled in the analyses.
Physiological Responses
Tables 1 and 2 show that the Time ⫻ Condition interaction
was not significant for cortisol or for any of the cardiovascular
Time ⫻
Sex main
Rosemary condit.
effect
interaction
Baseline Treatment F(1,23) p F(1,24) p
.188 (.117) .153 (.113) 0.53 .59 ⬍.01 .97
.181 (.056) .149 (.070)
601
 J. J. GEDNEY et al.
TABLE 2. Cardiovascular (CV) Measures (mm Hg and bpm) Obtained at Baseline and Immediately Posttreatment, Reported for Sex by
Treatment Condition

Time ⫻
Sex main
Control Lavender Rosemary condit.
effect
CV measure interaction

Baseline Treatment Baseline Treatment Baseline Treatment F(1,23) p F(1,24) p

SBP
Men 128.0 (12.5) 122.7 (11.6) 124.0 (14.9) 121.3 (12.1) 118.2 (13.0) 118.2 (13.0) 1.18 .32 10.46 ⬍.01
Women 109.0 (10.8) 108.8 (11.7) 107.5 (9.8) 107.6 (9.1) 107.6 (9.1) 107.4 (10.0)
DBP
Men 75.9 (13.2) 74.0 (12.8) 73.0 (10.8) 76.6 (10.8) 72.5 (10.4) 68.3 (9.4) 0.03 .97 0.69 .42
Women 72.4 (6.6) 73.2 (9.6) 70.5 (7.9) 70.5 (8.7) 71.1 (7.0) 74.1 (7.1)
HR
Men 71.2 (10.1) 69.2 (10.4) 66.7 (9.2) 68.0 (9.5) 71.0 (10.2) 69.2 (10.3) 0.07 .94 0.30 .59
Women 70.9 (7.2) 70.6 (10.0) 74.2 (11.0) 73.9 (10.7) 73.1 (8.5) 72.0 (8.7)

a
A significant main effect for sex was obtained for SBP. Standard deviations are reported in parentheses.
SBP ⫽ systolic blood pressure; DBP ⫽ diastolic blood pressure; HR ⫽ heart rate.

TABLE 3. Heat Threshold, Tolerance, and Pressure Pain Procedures Reported for Sex by Treatment Conditiona

Time ⫻
Sex main
Control Lavender Rosemary condit.
effect
Sensory test interaction

Pre-Tx Post-Tx Pre-Tx Post-Tx Pre-Tx Post-Tx F(1,23) p F(1,24) p

Heat threshold (°C)

Men 45.0 (2.0) 44.8 (1.6) 45.0 (2.8) 45.0 (2.6) 45.0 (2.4) 44.8 (2.4) 0.10 .91 2.26 .15
Women 43.7 (2.1) 43.8 (2.5) 44.0 (1.7) 43.7 (1.4) 43.9 (2.1) 44.1 (2.1)
Heat tolerance (°C)
Men 50.0 (1.3) 49.4 (1.4) 50.0 (1.5) 49.6 (1.3) 49.8 (1.2) 49.7 (1.3) 0.59 .56 6.69 .02
Women 48.5 (1.6) 48.2 (1.7) 48.6 (1.9) 48.3 (1.3) 48.7 (1.1) 48.2 (1.4)
Pressure, masseter (kg/cm2)
Men 3.6 (1.1) 3.5 (1.1) 3.5 (0.8) 3.6 (0.9) 3.1 (0.8) 3.1 (0.9) 0.17 .80 13.63 ⬍.01
Women 2.5 (0.6) 2.5 (0.7) 2.5 (0.7) 2.4 (0.8) 2.6 (0.6) 2.6 (0.8)
Pressure, trapezius (kg/cm2)
Men 7.8 (1.8) 7.2 (1.7) 7.6 (1.6) 7.4 (1.6) 6.3 (2.1) 6.2 (2.1) 0.23 .79 10.44 ⬍.01
Women 4.7 (1.6) 4.8 (1.5) 4.7 (1.7) 4.4 (1.7) 5.2 (1.8) 5.0 (2.1)

a
Significant main effects for sex were obtained for Heat Tolerance and Masseter and Trapezius Pressure. Significant main effects of time emerged for Heat
Tolerance and Pressure Pain Threshold at the Trapezius. Standard deviations are reported in parentheses.

measures (p values ⱖ .32), suggesting the absence of neu-
roendocrine or autonomic responses to treatment. With the
exception of systolic blood pressure, the main effect for sex
was likewise not significant, demonstrating that men and
women responded similarly and consistently within and be-
tween sessions (systolic blood pressure: p ⫽ .02, all others:
p ⬎ .31).
Experimental Pain Procedures
Tables 3 and 4 show that the Time ⫻ Condition interaction
was significant only for the 49°C Heat Pulse, calculated using
the pain rating change across the trial. Scores differed between
the Lavender and Rosemary conditions (p ⫽ .01), but not
between the Control and aroma treatment conditions. Signif-
icant main effects for sex were obtained for Heat Tolerance,
Masseter and Trapezius Pressure and 49°C and 52°C Heat
Pulse (maximum rating change across the trial) (p values ⱕ
.02). Consistently, women reported lower heat and pressure
tolerance and greater windup (heat pulse) pain relative to men.
Main effects of time emerged for several pain measures (Heat
Tolerance, Trapezius Pressure, Ischemic Threshold and 49°C
and 52°C Maximum Heat Ratings), with all effects except
Ischemic Threshold in the direction of greater pain sensitivity
during the postassessment.
Posttreatment Questionnaires
Table 5 and Figure 1 present data from the postsession
questionnaire. A significant difference in the strength of odor
between the control condition and each of the two essential
oils was reported (p values ⬍ .01). The Lavender treatment,
but not the Rosemary treatment was rated as marginally more
stimulating than the control condition (p values ⫽ .06 and
0.78, respectively). Retrospectively, subjects indicated greater
reductions in pain intensity and pain unpleasantness after
treatment with lavender, compared with the control condition
evaluations. Specifically, men reported diminished pain inten-

602 Psychosomatic Medicine 66:599 – 606 (2004)

AROMATHERAPY AND PAIN DISCRIMINATION
TABLE 4. Heat Pulse and Ischemic Pain Procedures Reported for Sex by Treatment Conditiona

Time ⫻
Sex main
Control Lavender Rosemary condit.
effect
Sensory test interaction

Pre-Tx Post-Tx Pre-Tx Post-Tx Pre-Tx Post-Tx F p F p

Heat pulse, 49°C max. rating

Men 24.7 (8.4) 26.1 (8.3) 27.3 (8.0) 30.5 (11.3) 25.7 (11.3) 23.8 (9.5) 5.10 .01 7.05 .01
Women 41.7 (23.5) 48.7 (24.5) 39.7 (24.0) 49.6 (26.5) 37.4 (16.5) 39.1 (16.9)
Heat pulse, 49°C rating change
Men 5.3 (5.3) 5.8 (4.9) 6.5 (4.3) 6.4 (6.3) 6.1 (3.6) 6.1 (4.4) 0.01 .99 2.16 .15
Women 10.7 (13.4) 9.1 (13.4) 9.8 (9.7) 9.3 (9.7) 11.3 (10.7) 10.5 (8.4)
Heat pulse, 52°C max. rating
Men 44.8 (18.6) 49.8 (24.2) 49.1 (19.8) 52.3 (18.5) 45.4 (19.7) 47.7 (21.4) 0.53 .59 9.37 .01
Women 73.4 (25.8) 79.0 (25.1) 71.3 (25.8) 74.9 (27.9) 69.2 (21.4) 78.3 (25.7)
Heat pulse, 52°C rating change
Men 12.9 (11.4) 15.5 (9.3) 19.2 (14.0) 19.4 (18.5) 18.8 (13.6) 18.2 (15.8) 0.36 .70 0.01 .93
Women 17.5 (15.9) 13.7 (13.4) 17.7 (20.0) 15.4 (16.9) 25.8 (13.3) 18.5 (13.9)
Ischemic, intensity
Men 12.0 (5.6) 11.4 (6.0) 10.9 (5.3) 11.0 (5.8) 12.0 (5.3) 11.3 (5.7) 1.16 .32 0.35 .56
Women 10.0 (5.4) 9.9 (5.9) 10.7 (5.6) 10.4 (5.7) 10.9 (4.9) 10.3 (5.6)
Ischemic, unpleasantness
Men 12.4 (5.3) 11.6 (5.7) 11.3 (4.9) 11.3 (5.6) 11.9 (4.8) 11.7 (5.4) 0.33 .72 0.03 .86
Women 11.5 (5.4) 11.4 (5.6) 12.3 (5.9) 12.0 (6.1) 11.7 (5.0) 11.6 (5.6)
Ischemic, threshold (seconds)
Men 141 (078) 193 (185) 164 (124) 202 (182) 150 (113) 190 (190) 1.88 .17 0.65 .43
Women 242 (237) 247 (240) 233 (252) 220 (236) 219 (193) 237 (228)
Ischemic, tolerance (seconds)
Men 611 (252) 659 (229) 705 (240) 712 (223) 651 (239) 709 (217) 1.06 .36 0.22 .64
Women 642 (266) 646 (273) 620 (276) 601 (264) 630 (288) 637 (288)

a
A significant time ⫻ condition interaction was obtained for heat pulse, 49°C (maximum rating), where pain rating differences were obtained between the
lavender and rosemary treatment conditions. Significant main effects for sex were obtained for heat pulse, 49°C and 52°C (maximum rating). Significant main
effects of time emerged for maximum ratings at 49°C and 52°C and for Ischemic Threshold. Standard deviations are reported in parentheses.

TABLE 5. Subjective Ratings of Aromatherapy Treatment, Reported by Treatment Condition: Paired-Samples T-Tests Reveal a Significant
Effect of Lavender (and Marginal Effect for Rosemary) on Perceived Intensity, and a Significant Effect of Lavender on Perceived Unpleasantness

Control– Lavender–
Control Lavender Rosemary Control–Rosemary
Lavender Rosemary
Odor ratinga
Mean Mean Mean
t pⱕ t pⱕ t pⱕ
(SD) (SD) (SD)

Strength of odor (0 ⫽ not noticeable, 7.0 (13.2) 61.3 (18.3) 65.9 (22.1) 11.5 .01 12.7 .01 0.7 .48
100 ⫽ strong)
Property of odor (0 ⫽ stimulating, 47.2 (7.9) 36.3 (26.9) 45.2 (24.5) 2.0 .06 0.3 .78 1.4 .17
100 ⫽ relaxing)
Odor effect on perceived pain intensity 51.3 (3.2) 56.2 (7.8) 55.8 (10.3) 2.9 .01b 2.0 .06 0.6 .53
(0 ⫽ increase, 100 ⫽ decrease)
Odor effect on perceived pain unpl. 50.8 (2.2) 56.4 (9.7) 54.1 (14.9) 2.7 .01c 1.1 .30 1.0 .31
(0 ⫽ increase, 100 ⫽ decrease)

a
Ratings based on a 100-mm visual analog scale. The midpoint (50 mm) is neutral for the last three rating scales. Standard deviations are reported in parentheses.
b
Men: Wilcoxon z ⫽ 2.14, p ⫽ .03; Women: Wilcoxon z ⫽ 1.68, p ⫽ .09.
c
Men: ns; women: Wilcoxon z ⫽ 2.08, p ⫽ .04.

sity (Wilcoxon z ⫽ 2.14, p ⫽ .03), whereas women reported and posttreatment psychophysical pain responses along with
diminished pain unpleasantness (Wilcoxon z ⫽ 2.08, p ⫽ .04). retrospective evaluations of analgesic effects were examined.

DISCUSSION Effect of Aromatherapy on Physiological Response

This study employed quantitative sensory testing to mea- The absence of pre- posttreatment differences indicates that
sure the analgesic effect of inhalation treatment using essential our treatment conditions did not elicit a detectable neuroen-
oil of lavender and essential oil of rosemary. Changes in pre- docrine or autonomic response. The absence of a cortisol sex

Psychosomatic Medicine 66:599 – 606 (2004) 603


Figure 1. Bar charts reporting subject ratings of odor strength and effect of treatment on pain intensity and pain unpleasantness. Significant differences were
obtained for pain intensity and pain unpleasantness between the Lavender and Control conditions, while a moderate difference for pain intensity was obtained
between the Rosemary and Control condition (Error bars ⫽ SEM).
main effect indicates that both men’s and women’s neuroen-
docrine responses were equal at the end of the 10-minute quiet
resting baseline and treatment periods. Though not signifi-
cantly different (with the exception of systolic blood pres-
sure), cardiovascular sex effects were in the expected direc-
tion.
Effect of Aromatherapy on Quantitative Sensory
Testing
The results of quantitative sensory testing revealed a sig-
nificant pre–post treatment difference across experimental
conditions only for the 49°C Heat Pulse task, when ratings
were calculated as the difference between the first and last of
the 10 pulses. It is noteworthy that this apparent treatment
effect was obtained between the Lavender and Rosemary
treatment conditions, and not between the Control and either
essential oil treatment condition. Thus, the absence of control
matched changes in pain response after inhalation of essential
oils indicates that the treatment condition we employed pro-
duced no analgesic effects. As reported by others (6), main sex
effects were observed for several of the pain measures (Heat
Tolerance, 49°C and 52°C Heat Pulse and Masseter and
Trapezius Pressure). All women in this study were scheduled
during the follicular phase of their menstrual cycle (4 –9 days
after the onset of menses). It is possible that controlling for
cycle phase eliminated potential sex differences for the other
pain measures (4). Also, all subjects in this study had partic-
ipated in a previous experimental pain investigation; therefore,
reduced novelty effects may have diminished any sex differ-
ences. Moreover, a selection bias may have attenuated sex
differences, because those subjects who found the pain stimuli
604
J. J. GEDNEY et al.
most distressing would have been less likely to volunteer for
another laboratory pain experiment.
To our knowledge, this is the first time that quantitative
sensory testing with healthy subjects has been used to test
for possible analgesic effects of inhaled essential oils. In
the only other published experimental pain report, Gobel et
al. topically administered four different test preparations to
the forehead and temples of 32 healthy males, using a
double-blind, randomized, crossover design (12). Test
preparations included combinations of peppermint oil, eu-
calyptus, and ethanol. Pre- and postmanipulation measures
included temporal EMG activity, mood state, and sensitiv-
ity to experimental pain. Pain stimuli were selected to simulate
tension-type headache symptoms. Stimuli included ischemic
pain (pressure cuff placed around the cranium), mechanical
pressure pain (pressure algometer applied to the scalp and
middle phalanx of the middle finger), and thermal pain (con-
tact thermistor applied to the forehead). After treatment with
peppermint oil, subjects reported decreased emotional irrita-
tion and demonstrated decreased resting EMG. Ischemic pain,
but not thermal or mechanical pain ratings were significantly
reduced. It is possible that the topical application, versus the
inhalation used in our study, was responsible for the pain-
reducing effects of their intervention. It remains to be seen
how the topical application of essential oils may affect quan-
titative sensory testing.
Retrospective Reports of Treatment Effects
Retrospective evaluation of the effects of aromatherapy
suggest that when asked to self-report any perceived change in
pain sensitivity after treatment, subjects reported that they
Psychosomatic Medicine 66:599 – 606 (2004)
AROMATHERAPY AND PAIN DISCRIMINATION
experienced less pain intensity and pain unpleasantness after
the Lavender treatment and a trend toward less pain intensity
after the Rosemary treatment, relative to the Control condi-
tion. There was no clear sex-related trend in this self-report,
although men reported diminished pain intensity and women
reported diminished pain unpleasantness in the Lavender con-
dition. Although the inhalation of the essential oils did not
elicit a detectable analgesic response, the use of these oils may
have provided an additional pleasant sensory and affective
stimulus, which could alter overall evaluation of the pain
experience, thus influencing retrospective evaluation of the
treatment effects. This interpretation is consistent with those
of others who have demonstrated aroma-mediated changes
in affect in both the clinical and laboratory setting
(4 – 6,9,18,22,24,38 – 40). The role of affectivity as an in-
tegral component of the pain experience has been well
documented (41– 47). From the results of the present study,
it appears that the use of aromatherapy for the management
of acute pain may produce beneficial effects not through
direct physiological mechanisms, but by providing a com-
peting pleasant sensory and affective experience that can
alter retrospective pain evaluation.
Limitations
Several limitations of this study deserve mention. First,
subjects in this study were familiar with sensory testing meth-
ods through previous participation in another sensory study,
which is both a strength and a limitation. The benefit is that
the variability associated with novelty effects was reduced;
however, recruiting experienced subjects may have introduced
a selection bias, and a different pattern of results may have
emerged with naı̈ve participants. Another limitation is the
relatively small sample size; however, given the minimal
effect sizes, the sample size required to detect significant
aromatherapy effects on psychophysical responses would be
prohibitive. In addition, the aromatherapy intervention may
not have been potent enough to alter pain responses, and a
more intense and/or more prolonged aromatherapy treatment
may have been more effective. Lastly, it is possible that
providing odor property ratings (ie, analog rating from stim-
ulating to relaxing) immediately before the posttreatment pain
ratings may have biased a response toward diminished pain
ratings. However, this seems unlikely as the subjects’ affec-
tive response to the odor was established before the posttreat-
ment pain stimuli and postsession questionnaire.
CONCLUSION
In this study, we demonstrated that inhalation of essential
oil of lavender and essential oil of rosemary does not produce
a detectable analgesic effect; however, subjects’ retrospective
evaluations of aroma-induced changes in pain intensity and
pain unpleasantness suggested that they perceived some ben-
efit of the intervention, especially for lavender. Thus, in a
typical clinical evaluation protocol that relies on retrospective
evaluations of treatment effectiveness, aromatherapy may pro-
duce a clinically relevant shift in the patient’s report of per-
Psychosomatic Medicine 66:599 – 606 (2004)
ceived pain. Therefore, given the minimal side effects of this
intervention, aromatherapy may be helpful in treatment set-
tings associated with both pain and heightened affective
arousal, such as dental care or the outpatient treatment setting.
REFERENCES
1. Buckle J, Clinical aromatherapy in nursing. 2nd ed. New York: Arnold
Publishers; 1997.
2. Ehrlichman H, Halpern JN. Affect and memory: effects of pleasant and
unpleasant odors on retrieval of happy and unhappy memories. J Pers Soc
Psychol 1988;55:769 –79.
3. Degel J, Koster EP. Odors: implicit memory and performance effects.
Chem Senses 1999;24:317–25.
4. Lehrer J, Eckersberger C, Walla P, Potsch G, Deecke L. Ambient odor of
orange in a dental office reduces anxiety and mood in female patients.
Physiol Behav 2000;71:83– 6.
5. Dale A, Cornwell S. The role of lavender oil in relieving perineal
discomfort following childbirth: a blind randomized clinical trial. J Adv
Nurs 1994;19:89 –96.
6. Marchand S, Arsenault P. Odors modulate pain perception: a gender-
specific effect. Physiol Behav 2002;76:251– 6.
7. Ilmberger J, Heuberger E, Mahrhofer C, Dessovic H, Kowarik D, Buch-
bauer G. The influence of essential oils on human attention. I: alertness.
Chem Senses 2001;26:239 – 45.
8. Buckle J. Use of aromatherapy as a complementary treatment for chronic
pain. Alternative Ther 1999;5:42–51.
9. Diego MA, Jones NA, Field T, Hernandez-Reif M, Schanberg S, Kuhn C,
McAdam V, Galamaga R, Galamaga M. Aromatherapy positively affects
mood, EEG patterns of alertness and math computations. Int J Neurosci
1998;96:217–24.
10. Sparks MJ, O’Sullivan P, Herrington AA, Morcos SK. Does peppermint
oil relieve spasm during barium enema? Br J Radiol 1995;68:841–3.
11. Gobel H, Fresenius J, Heinze A, Dworschak M, Soyka D. Effectiveness
of Oleum menthae piperitae and paracetamol in therapy of headache of
the tension type. Nervenarzt 1996;67:672– 81.
12. Gobel H, Schmidt G, Soyka D. Effect of peppermint and eucalyptus oil
preparations on neurophysiological and experimental algesimetric head-
ache parameters. Cephalalgia 1994;14:228 –34.
13. Bengtsson S, Berglund H, Gulyas B, Cohen E, Savic I. Brain activation
during odor perception in males and females. Neuroreport 2001;12:
2027–33.
14. Hardy M, Kirk-Smith MD, Stretch DD. Replacement of drug treatment
for insomnia using ambient odour. Lancet 1995;346:701.
15. Liu JH, Chen GH, Yeh HZ, Huang CK, Poon SK. Enteric-coated pep-
permint-oil capsules in the treatment of irritable bowel syndrome: a
prospective, randomized trial. J Gastroenterol 1997;32:765– 8.
16. Saeki Y, Shiohara M. Physiological effects of inhaling fragrances. Int J
Aromather 2001;11:118 –25.
17. Saeki Y. The effect of foot-bath with or without the essential oil of
lavender on the autonomic nervous system: a randomized trial. Comple-
ment Ther Med 2000;8:2–7.
18. Edge J. A pilot study addressing the effect of aromatherapy massage on
mood, anxiety and relaxation in adult mental health. Complement Ther
Nurs Midwifery 2003;9:90 –7.
19. Lis-Balchin M, Hart S, Wan Hang, Lo B. Jasmine absolute (Jasminum
grandiflora L.) and its mode of action on guinea-pig ileum in vitro.
Phytother Res 2002;16:437–9.
20. Buchbauer G, Jirovetz L, Jager W, Plank C, Dietrich H. Fragrance
compounds and essential oils with sedative effects upon inhalation.
J Pharm Sci 1993;82:660 – 4.
21. Woolfson A, Hewitt D. Intensive aromacare. Int J Aromather 1992;4:
12– 4.
22. Buckle J. Does it matter which lavender essential oil is used? Nurs Times
1993;89:32–5.
23. Moss M, Cook J, Wesnes K, Duckett P. Aromas of rosemary and
lavender essential oils differentially affect cognition and mood in healthy
adults. Int J Neurosci 2003;113:15–38.
24. Ching M. Contemporary therapy: aromatherapy in the management of
acute pain? Contemp Nurse 1999;8:146 –51.
25. Edwards RR, Fillingim RB. Age-associated differences in responses to
noxious stimuli. J Gerontol A Biol Sci Med Sci 2001;56:M180 –5.
26. Robinson ME, Riley JL 3rd, Myers CD, Papas RK, Wise EA, Waxenberg
605

LB, Fillingim RB. Gender role expectations of pain: relationship to sex
differences in pain. J Pain 2001;2:251–7.
27. Myers CD, Robinson ME, Riley JL 3rd, Sheffield D. Sex, gender, and
blood pressure: contributions to experimental pain report. Psychosom
Med 2001;63:545–50.
28. Wise EA, Price DD, Myers CD, Heft MW, Robinson ME. Gender role
expectations of pain: relationship to experimental pain perception. Pain
2002;96:335– 42.
29. Yousem DM, Maldjian JA, Siddiqi F, Hummel T, Alsop DC, Geckle RJ,
Bilker WB, Doty RL. Gender effects on odor-stimulated functional
magnetic resonance imaging. Brain Res 1999;818:480 –7.
30. Fillingim RB, Maixner W, Girdler SS, Light KC, Harris MB, Sheps DS,
Mason GA. Ischemic but not thermal pain sensitivity varies across the
menstrual cycle. Psychosom Med 1997;59:512–20.
31. Pfleeger M, Straneva PA, Fillingim RB, Maixner W, Girdler SS. Men-
strual cycle, blood pressure and ischemic pain sensitivity in women: a
preliminary investigation. Int J Psychophysiol 1997;27:161– 6.
32. Watson D, Clark LA, Tellegen A. Development and validation of brief
measures of positive and negative affect: the PANAS scales. J Pers Soc
Psychol 1988;54:1063–70.
33. Spielberger CD, Gorusch R, Lushene RE, Vagg PR, Jacobs GA. Manual
for the State-Trait Anxiety Inventory (Form Y1). Palo Alto, CA: Con-
sulting Psychologists Press; 1983.
34. McNair D, Lorr M, Droppleman LF. Profile of mood states manual. San
Diego, CA: Educational and Industrial Testing Service; 1971.
35. Moore PA, Duncan GH, Scott DS, Gregg JM, Ghia JN. The submaximal
effort tourniquet test: its use in evaluating experimental and chronic pain.
Pain 1979;6:375– 82.
36. Gracely RH, Kwilosz DM. The Descriptor Differential Scale: applying
psychophysical principles to clinical pain assessment. Pain 1988;35:
279 – 88.
37. Kirschbaum C, Hellhammer DH. Salivary cortisol in psychoneuroendo-
606
J. J. GEDNEY et al.
crine research: recent developments and applications. Psychoneuroendo-
crinology 1994;19:313–33.
38. Baron RA. Environmentally induced positive affect: its impact on self-
efficacy, task performance, negotiation, and conflict. J Appl Soc Psychol
1990;20:368 – 84.
39. Robin O, Alaoui-Ismaili O, Dittmar A, Vernet-Maury E. Basic emotions
evoked by eugenol odor differ according to the dental experience. A
neurovegetative analysis. Chem Senses 1999;24:327–35.
40. Wheeler-Robins JL. The science and art of aromatherapy. J Holist Nurs
1999;17:5–17.
41. Price DD, Riley JL 3rd, Wade JB. Psychophysical approaches to the
measurement of the dimensions and stages of pain. In: Melzack R, Turk
DC, editors. Handbook of pain measurement. New York, NY: The
Guilford Press; 2001. p. 53–75.
42. Gedney JJ, Logan HL, Baron RS. Predictors of short-term and long-term
memory of sensory and affective dimensions of pain. J Pain 2003;4:
47–55.
43. Logan HL, Gedney JJ, Sheffield D, Xiang Y, Starrenberg E. Stress
influences the level of negative affectivity following forehead cold pres-
sor pain. J Pain 2003;4:520 –9.
44. Diagnostic criteria for functional gastrointestinal disorders. In: Drossman
D, Corazziari E, Talley N, Thompson W, Whitehead WE, editors. The
functional gastrointestinal disorders. Lawrence, KS: Allen Press; 2000. p.
659 – 69.
45. Sheffield D, Logan H, Lutgendorf S, Haldeman R, Iverson K, Wade R,
Kirchner HL, Lubaroff D. Effects of stress and relaxation on capsaicin-
induced pain and flare response. Br J Health Psychol 2003;9 (in press).
46. Price DD. Psychological and neural mechanisms of the affective dimen-
sion of pain. Science 2000;288:1769 –72.
47. Janssen SA. Negative affect and sensitization to pain. Scand J Psychol
2002;43:131–7.
Psychosomatic Medicine 66:599 – 606 (2004)