Shen J, Wenger N, Glaspy J, Hays RD, Albert PS, Choi C, Shekelle PG. Electroacupuncture for Control of Myeloablative Chemotherapy–Induced EmesisA Randomized Controlled Trial. JAMA. 2000;284(21):2755–2761. doi:10.1001/jama.284.21.2755
Author Affiliations: Laboratory of Clinical Studies, Brain Electrophysiology and Imaging Section, National Institute of Alcoholism and Alcohol Abuse, National Institutes of Health, Bethesda, Md (Dr Shen); Department of Medicine, Division of General Internal Medicine and Health Services Research (Drs Wenger and Hays) and Department of Medicine, Division of Hematology and Oncology (Dr Glaspy), UCLA School of Medicine, Los Angeles, Calif; RAND, Santa Monica, Calif (Drs Hays and Shekelle); Biometric Research Branch, National Cancer Institute, National Institutes of Health, Bethesda, Md (Dr Albert); and Department of Medicine, West-L.A. VAMC, Veterans Affairs Health Services Research and Development Service, Los Angeles, Calif (Dr Shekelle). Dr Choi is in private practice in Santa Monica, Calif.
Context High-dose chemotherapy poses considerable challenges to emesis management.
Although prior studies suggest that acupuncture may reduce nausea and emesis,
it is unclear whether such benefit comes from the nonspecific effects of attention
and clinician-patient interaction.
Objective To compare the effectiveness of electroacupuncture vs minimal needling
and mock electrical stimulation or antiemetic medications alone in controlling
emesis among patients undergoing a highly emetogenic chemotherapy regimen.
Design Three-arm, parallel-group, randomized controlled trial conducted from
March 1996 to December 1997, with a 5-day study period and a 9-day follow-up.
Setting Oncology center at a university medical center.
Patients One hundred four women (mean age, 46 years) with high-risk breast cancer.
Interventions Patients were randomly assigned to receive low-frequency electroacupuncture
at classic antiemetic acupuncture points once daily for 5 days (n = 37); minimal
needling at control points with mock electrostimulation on the same schedule
(n = 33); or no adjunct needling (n = 34). All patients received concurrent
triple antiemetic pharmacotherapy and high-dose chemotherapy (cyclophosphamide,
cisplatin, and carmustine).
Main Outcome Measures Total number of emesis episodes occurring during the 5-day study period
and the proportion of emesis-free days, compared among the 3 groups.
Results The number of emesis episodes occurring during the 5 days was lower
for patients receiving electroacupuncture compared with those receiving minimal
needling or pharmacotherapy alone (median number of episodes, 5, 10, and 15,
respectively; P<.001). The electroacupuncture
group had fewer episodes of emesis than the minimal needling group (P<.001), whereas the minimal needling group had fewer
episodes of emesis than the antiemetic pharmacotherapy alone group (P = .01). The differences among groups were not significant
during the 9-day follow-up period (P = .18).
Conclusions In this study of patients with breast cancer receiving high-dose chemotherapy,
adjunct electroacupuncture was more effective in controlling emesis than minimal
needling or antiemetic pharmacotherapy alone, although the observed effect
had limited duration.
High-dose, multiple-day, multiple-drug myeloablative chemotherapy poses
substantial challenges to emesis control. The combination of chemotherapy
agents is highly emetogenic; most patients have experienced emesis with multiple
courses of prior chemotherapy; and patients may have received other medical
care or medications and adjuncts that can contribute to emesis. During the
last 2 decades, new effective antiemetic pharmacological agents have helped
to improve control of chemotherapy-induced emesis. Because of concerns about
pharmacokinetic interaction between high-dose chemotherapy agents and the
new antiemetic medications, some patients receiving intense multiple-agent,
myeloablative chemotherapy regimens might not be able to use these newer antiemetics
This constellation of factors makes the management of emesis difficult.
A recent National Institutes of Health Consensus Development Conference
report concluded that acupuncture was efficacious in reducing emesis associated
with chemotherapy.4 However, it was suspected
that such benefit was due to a placebo effect. The one published randomized
controlled trial that used a sham acupuncture control included 10 patients
and was limited by its use of a crossover design without a washout period.5 Other data that support the use of acupuncture to
control emesis include a systematic review and a recent meta-analysis in the
postoperative setting.6,7 However,
physiologic reasons suggest that successful treatment with acupuncture might
be different between chemotherapy–induced and postoperative emesis.8,9
In this study, we assessed a standardized electroacupuncture protocol
as an adjunct to antiemetic pharmacotherapy for controlling emesis associated
with intensive, multiple-drug, combination myeloablative chemotherapy compared
with minimal needling or antiemetic pharmacotherapy alone. We hypothesized
a priori that minimal needling treatment might have a greater effect than
pharmacotherapy alone because of the nonspecific effects of needles and the
additional attention and care the patients received.
This study was conducted at a tertiary teaching hospital with a comprehensive
cancer center. Patients were recruited from oncology clinics between March
1996 and December 1997 and enrolled successively after written informed consent
was obtained. The assessment and intervention procedures were administered
when the patients were hospitalized in oncology wards for myeloablative chemotherapy.
The study protocol was approved by the local cancer center scientific peer
review committee and the institutional human subject protection committee.
Female patients 18 to 62 years of age were eligible if they had histologically
proven resected breast cancer, Karnofsky performance status greater than 80
(on 0-100 scale), life expectancy of at least 6 months, and were appropriate
candidates for the bone marrow transplantation program. We excluded patients
who had brain metastases; life's threatening concurrent nonmalignant conditions;
active infection, including an active skin infection over the proposed treatment
area; any condition that compromised their ability to give informed consent;
or a cardiac pacemaker. A research assistant conducted a face-to-face interview
to obtain the following information: sociodemographic characteristics, history
of chemotherapy, history of nausea and emesis related to motion sickness,
morning sickness, patients' expectations about the adverse effects of chemotherapy,
and the benefits of antiemetic treatment.
The patients were randomly assigned without stratification to receive
1 of 3 treatment options: electroacupuncture, minimal needling, or antiemetic
drugs alone. The 2 interventions were described to the patient as "classical
acupuncture" or "non-classical acupuncture." We used the term classical acupuncture to describe a protocol for electroacupuncture
at sites that are indicated for nausea and emesis control. We used the term non-classical acupuncture to describe a protocol of minimal
needling near sites that are not indicated for nausea and emesis control with
mock stimulation. Patients were informed that the beneficial effect of either
treatment is not known.
Serially numbered, sealed, opaque envelopes were used to indicate assignment.
An investigator who had no direct contact with the study patients prepared
the envelopes, using a random number table to generate the sequence. Patients
were entered into a study log before the envelopes were opened. All envelopes
were accounted for.
Chemotherapy. All patients in the 3 groups received the same chemotherapy regimen
and antiemetic drugs, following a standard protocol: On hospital days 1, 2,
and 3, all patients received high doses of cyclophosphamide and cisplatin,
and on day 4, carmustine. Administration of the chemotherapy drugs were as
follows: cyclophosphamide, 1875 mg/m2 body-surface area, per day
over 60 minutes, starting at 9 AM for 3 days; cisplatin, 55 mg/m2
body-surface area, per day with continuous infusion, starting at 9 AM for
3 days; and on day 4, carmustine, 600 mg/m2 body-surface area,
over 2 hours, starting at 9 AM, immediately after the cisplatin dose was completed.
Antiemetic Agents. All patients also received the same triple pharmacological agents for
emesis management. The regimen included prochlorperazine prechemotherapy loading,
10 mg in 100 mL of normal saline, intraveneously, over 10 minutes, followed
by continuous intravenous infusion at 1 mg/m2 body-surface area
per hour; lorazepam, 1 mg/m2 body-surface area, intravenously,
every 4 hours; and diphenhydramine hydrochloride, 25 mg/m2 body-surface
area, intravenously, every 6 hours. These medications were started 1 hour
prior to chemotherapy and were continued until 48 hours after the last chemotherapy
infusion. Rescue medications that were available to all patients included
additional protocol agent prochlorperazine, lorazepam, and diphenhydramine,
as well as metoclopramide and droperidol. Additional medications were given
at the discretion of staff physicians not involved in this study as well as
at the patient's request.
Electroacupuncture. Patients in the electroacupuncture group received adjunct treatment
that consisted of perpendicular insertion of a 36-gauge disposable stainless
steel acupuncture needle (Seirin, Japan) at PC6 acupuncture point, located
between the tendons of palmaris longus and flexor carpii radialis at 2 body-inches
(a body-inch or a cun is the greatest width of a
patient's thumb at the distal phalanx) above the wrist crease. (This article
uses the international nomenclature agreed on by the World Health Organization
in 198910). The depth of insertion was 1 body-inch.
The needle was inserted with bilateral rotation without introducer and was
manipulated until achieving a "de Qi" sensation, that is, the acupuncturist
feels sensations from the needle-manipulation and the patient feels soreness,
fullness, heaviness, or local area distention. The needling technique included
twirling, thrusting, lifting, and initial flicking. After de Qi was achieved,
the needle was connected through a microalligator clip and an electrode to
a battery-operated pulse generator connected to the negative pole. A second
needle was inserted perpendicularly to a depth of 1.5 body-inches at the ST36
acupuncture point (located between the tibialis anterior muscle and the tendon
of the extensor digitorum longus pedis at 1-finger's breadth lateral to the
lower border of tibia tuberosity) with a microalligator clip and an electrode
connected to the positive pole. The needling technique included twirling,
thrusting, lifting, initial flicking, and periosteal pecking. These procedures
were performed bilaterally. This acupunture protocol was based on prior literature,
acupuncture textbooks, and suggestions from consultant practitioners following
a common symptomatic approach.11
Electrical frequency was delivered over 2 to10 Hz, 0.5 to 0.7 milliseconds
duration pulse width, under a variable direct current output with square waveform
balanced alternating polarity of less than 26 mA for 20 minutes (maximal voltage
15 V). Needling sites were examined at the end of each treatment. Two acupuncture
consultants observed the treatment procedure at the start of the study and
confirmed the technique.
Two investigators (J.G. and C.C.)administered the procedure in collaboration.
One was a clinical instructor at the medical school and had 3 years of acupuncture
training; the other was an acupuncture clinician with 20 years of practicing
experience. Before the needles were inserted, the clinicians evaluated patients
according to traditional pulse diagnosis procedures. The diagnostic ritual
was followed daily, although only the standard electroacupuncture protocol
was administered. The evaluation and treatment procedure took 30 minutes.
The first treatment was scheduled to occur within the 2 hours before the initial
chemotherapy infusion, usually between 7 AM and 9 AM on hospital day 1. Treatment
was given at the same time over the ensuing days for a total of 5 treatments.
Minimal Needling. For patients in the minimal needling group, a 36-gauge disposable acupuncture
needle was inserted subcutaneously with no manipulation or stimulation near
the LU7 acupuncture point, located on the lateral aspect of the radius proximal
to the styloid process. Care was taken to avoid de Qi sensation. The needle
was then connected through a microalligator clip and an electrode to a battery-operated
stimulator. A second needle was inserted subcutaneously, with no manipulation
or stimulation, near the GB34 acupuncture point, located in the depression
anterior and inferior to the head of the fibula, then connected to the stimulator.
The needle insertion procedure was performed bilaterally. One acupuncture
consultant observed the treatment procedure at the start of the study and
confirmed the technique.
This minimal needling protocol was designed by 2 traditional Chinese
medical physicians (C.C.). In their opinion, the protocol would not harm patients
and, at best, it could improve patients' "lung" and "muscle-skeletal" conditions.
However, they thought that the procedure was unlikely to prevent emesis. The
stimulator delivered the same audiovisual stimuli as in the electroacupuncture
technique for 20 minutes, but no electrical current
was passed to the needles. Needling sites were examined at the end of each
Treatment was administered by the same clinicians who administered the
electroacupuncture. The diagnostic and treatment ritual was the same as that
received by the electroacupuncture group, following the same schedule.
Pharmacotherapy. Patients in the pharmacotherapy alone control group received triple
pharmacotherapy agents described above and daily morning visits by physicians
and nurses but received no adjunct acupuncture or minimal needling therapy.
The primary outcome was the total number of emesis episodes that occurred
during the 5-day study period. Emesis was defined
as projection of gastric contents with resultant emesis product. Severe retching
without projection of gastric contents was not considered as an emesis episode.
The nurse's recording of the number of daily emesis episodes was used to measure
emesis. Nurses were not informed of the treatment group to which a patient
was assigned. A count of zero on a study day meant that the patient experienced
no emesis. As a secondary outcome, we compared the proportion of emesis-free
days across the 3 treatment groups.
For the primary outcome, we summed the daily counts over the 5-day study
period. We also instructed patients to record any adverse events that they
thought might be attributed to the study. After the 5-day study period, we
followed up patients for an additional 9 days, for a total of 2 weeks (14
We abstracted the number of emesis episodes from the nursing records
along with strict daily input and output records as measured in milliliters
(data not reported here). We also abstracted data from medical charts on the
concurrent antiemetic medications administered. Data were collected and entered
into electronic files by research assistants who had no knowledge of a patient's
treatment group assignment.
At the end of the 5-day study period, we asked patients who received
either type of adjunct needling treatment to complete a questionnaire evaluating
various aspects of the procedures. Questions included ratings of the technical
quality of the treatment, friendliness of the physician administering the
treatment, and the comfort level of the procedure. In addition, we asked patients
the treatment group to which they thought they were assigned.
Sample size and power estimation were based on 2-group comparisons of
emesis counts, assuming the mean count over the study period was 15 with Poisson
distribution (variance equal to the mean). A design with 35 patients in each
of the 3 groups will have 93% and 77% power, respectively, to detect a reduction
in the mean emesis count of 25% and 20% with a 2-sided test at the α
level of .017 (adjusting for the 3 pairwise comparisons, .05/3). Such effect
sizes are conservative estimates based on prior literature.12,13
We analyzed data according to the intention-to-treat principal, that
is, based on all randomized patients, as randomized. For the primary analysis,
the unit of analysis was the patient. We summed the number of emesis episodes
over 5 days14 and analyzed the total as a count
variable. Because the primary outcome was not normally distributed and highly
variable (ie, some patients had a substantial number of emesis episodes),
we analyzed the data using nonparametric tests and quasi-likelihood/Poisson
We first compared the nonadjusted outcome among 3 treatment groups using
the Kruskal-Wallis test followed by the Wilcoxon rank-sum test for pairwise
comparisons. Then, to adjust for a common set of baseline variables predictive
of the number of emesis episodes and to take into account the highly skewed
distribution of the emesis data, we conducted multivariate analyses using
quasi-likelihood/Poisson models.15 These baseline
variables included the patient's age, emesis experience with previous chemotherapy,
alcohol use, and experience with anticipatory nausea prior to chemotherapy.
The variable selection was based on prior literature16- 18
and achieving a parsimonious model. The Wald test was used to test for differences
between treatment groups. For all pairwise comparisons of the 3 treatment
groups, we considered a P = .05 to indicate statistical
significance. According to the Bonferroni adjustment, we used P<.017 to adjust for multiple comparison of 3 pairwise comparisons.
All P values were 2-tailed.
As a secondary analysis, we compared the proportion of emesis-free days
across the 3 treatment groups, adjusting for the same set of baseline variables
as in the primary analysis. This analysis was done using a generalized estimating
equation (GEE)/logistic model with robust variance estimation.19
Additional secondary analyses were performed to examine the effect of treatment
during the follow-up period; the outcome was the total number of emesis episodes
that occurred during the follow-up period, days 6 through 14. First, we used
the Kruskal-Wallis test followed by the Wilcoxon rank-sum test to test for
differences in the frequency of emesis episodes among groups during the follow-up
period. Second, GEE/ Poisson models were used to examine whether the effect
of treatment on the frequency of emesis episodes was diminished during the
follow-up period. We adjusted for the same set of baseline variables in this
regression model as we did in the primary analysis. Third, GEE/logistic models
were used to examine for a diminished effect, treat the outcomes as either
the presence or absence of emesis episodes on a given day, with and without
adjusting for the same set of baseline variables.
The recruitment and follow-up of study patients is shown in Figure 1. We recruited 111 consecutive eligible
patients, of whom 104 were randomized (2 refused randomization; 4 reported
a fear of acupuncture needles; and 1 developed hemothorax from a surgical
catheter before randomization). Thirty-seven patients were assigned to receive
electroacupuncture, 33 to minimal needling, and 34 to pharmacotherapy alone.
All but 1 patient received her intended chemotherapy treatments (1 individual
received only a partial dose of carmustine due to intolerance) and 2 patients
partly deviated from their adjunct minimal needling or electroacupuncture
treatment protocol (1 patient was transferred to a cardiac care unit for cardiac
toxicity and did not receive all intended intervention sessions; 1 individual
missed a treatment session because of a surgical procedure to replace a central
Two patients experienced adverse effects as a result of the electroacupuncture
or minimal needling procedure: 1 patient complained of an electrical shock
sensation from the needle-and-stimulator apparatus at the end of the first
treatment session and the equipment was immediately removed. The patient reported
no complaints on subsequent treatment days. One patient, who had residual
peripheral neuropathy manifested as tingling and numbness from prior chemotherapy,
complained of an aggravated tingling sensation following each needling procedure.
The mean age of participants was 46 years. About two thirds of the patients
reported their ethnicity as white and most had a college education (Table 1). All patients previously had received
chemotherapy, a mean of 6 courses. The vast majority had experienced nausea
with prior chemotherapy, and two thirds had experienced emesis associated
with previous chemotherapy. More than one fourth reported anticipatory nausea
associated with chemotherapy. Patients' baseline characteristics were well
matched among groups, except that patients assigned to receive minimal needling
had significantly more emesis with prior chemotherapy (difference among groups, P = .01).
Most patients reported that they did not know the group to which they
were assigned (between patients in the electroacupuncture group and those
in the minimal needling group: Pearson χ22 = 4.38; P = .11) (Table 2).
Patients' ratings of the clinician administering the treatment, the clinician's
friendliness, and the technical quality of the treatment they received were
comparable across the groups. The majority of the patients gave ratings of
excellent, and there were no statistical differences in ratings between those
in the electroacupuncture group and those in the minimal needling group. On
a 0- to 10-point scale, with 0 being not comfortable at all and 10 being very
comfortable, patients in the minimal needling group had a mean rating of 9.1
for comfort of their treatment, compared with a mean rating of 7.8 in the
electroacupuncture procedure group (Wilcoxon rank-sum test; P = .13).
All patients received antiemetic medications. Overall, when adjusted
to a 24-hour period, a patient received a mean dose of lorazepam, 9.8 mg;
diphenhydramine, 103.3 mg; and prochlorperazine, 42.4 mg, per day, including
the protocol and rescue doses, during the 5-day study period. The mean amount
of triple antiemetic agents administered during the study period was similar
across the 3 groups (Table 3).
During the same period, the use of rescue antiemetics administered in addition
to the triple agents also was comparable (Pearson χ22 = 2.23; P = .33; data not shown).
The study outcomes are shown in Table
4. Total emesis episodes per person over the 5-day study period
differed among the 3 groups (Kruskal-Wallis rank test; P<.001) (Figure 2). Pairwise
comparisons between groups showed that the electroacupuncture group had significantly
fewer emesis episodes than the minimal needling group (Wilcoxon rank-sum test; P<.001) or the pharmacotherapy alone group (Wilcoxon
rank-sum test; P<.001), and the minimal needling
group had significantly fewer emesis episodes than the pharmacotherapy alone
group (Wilcoxon rank-sum test; P = .01). Using a
quasi-likelihood/Poisson model to adjust for potentially confounding factors,
including age, alcohol use, emesis experience with prior chemotherapy, and
anticipatory nausea, the differences among groups remained significant. Pairwise
comparisons showed that both the minimal needling and electroacupuncture groups
had fewer emesis episodes than the pharmacotherapy alone group (β coefficient
for electroacupuncture group = − 0.36; SE, 0.068; P< .001; and β coefficient for minimal needling group = −
0.12; SE, 0.063; P = .02). Furthermore, the electroacupuncture
group had fewer emesis episodes than the minimal needling group (the difference
in the β coefficients for electroacupuncture group minus minimal needling
group = − 0.24; SE, 0.073; P<.001).
As a secondary analysis, patients in the electroacupuncture group had
a greater proportion of emesis-free days than patients in either the minimal
needling group or the pharmacotherapy group (Kruskal-Wallis rank test, P<.001). A GEE/logistic model adjusting for the same
baseline variables as in the primary analysis showed that the proportion of
emesis-free days was significantly greater in the electroacupuncture group
compared with either the minimal needling group (P<.001) or the pharmacotherapy
alone group (P<.001). The proportion of emesis-free
days was not significantly lower in the minimal needling group as compared
with the pharmacotherapy alone group (P = .18).
We also used 3 sets of analyses to address the issue of whether the
treatment effect from the study period diminished during the follow-up period.
Overall, comparing all 3 groups, the sums of emesis episodes as well as the
proportion of emesis-free days during days 6 to 14 were not significantly
different (Kruskal-Wallis rand test, P = .18 and P = .39, respectively) We used a GEE/Poisson regression
model to test whether the treatment effect was diminished during the follow-up
period while adjusting for the same set of baseline variables as in our primary
analysis. Results revealed that the difference in emesis counts between the
pharmacotherapy and the electroacupuncture groups was diminished in the follow-up
period (P<.001). All other differences (electroacupuncture
group vs minimal needling group and minimal needling group vs pharmacotherapy
alone group) were also diminished and were not statistically significant.
Similar results were found using the GEE-binary models to treat the outcome
as either the presence or absence of an episode on a given day.
In summary, during the follow-up period days 6 to 14, when groups were
no longer receiving adjunctive therapy, there were no significant differences
among groups in the number of emesis episodes or proportion of emesis-free
The results of this study suggest that the addition of daily electroacupuncture
treatment to this antiemetic regimen was superior to the pharmacotherapy therapy
alone or minimal needling in preventing chemotherapy–induced emesis.
There also was a trend indicating that the minimal needling procedure itself
was more effective in reducing emesis episodes than pharmacotherapy alone.
The observed differences between groups diminished in the follow-up period,
which further supported an antiemetic effect of electroacupuncture.
Our study has several strengths and limitations. A principal strength
is our use of a minimal needling intervention to assess the potential nonspecific
effects of needling and attention and care for the patient. While we could
not mask the acupuncturists to the interventions they were delivering, we
were able to achieve adequate masking of the patients. This strengthens the
internal validity of our study and increases the likelihood that the effect
we observed was due to a specific effect of electroacupuncture. A second strength
is our use of an electroacupuncture protocol that was standardized, reproducible,
and in agreement with the practices of our consultant practitioners, making
it more likely that others can attempt to replicate our results. Such attempts
at replication are needed as we used only 2 acupuncturists working at a single
The homogeneity of the patient population and their receipt of a standard
protocol for chemotherapy and supportive care increased the precision with
which we could measure the treatment effects but limits the generalizability
of our findings to other patient populations or those receiving other adjunct
therapies. The antiemetic pharmacotherapy used in our study protocol does
not include corticosteroids or a serotonin antagonist, such as ondansetron.20 Use of these agents has been shown to be superior
to the agents used in this study for controlling emesis. The effect of electroacupuncture
as an adjunct to other antiemetic regimens, including serotonin antagonists
and corticosteriods, is unknown.
It is important to note that minimal needling led to a reduction in
the frequency of emesis episodes. Attention and the clinician-patient interaction
are possible explanations for the beneficial effect in this setting. This
finding supports a role for behavioral interventions concomitant to pharmacological
management and also suggests that, in future studies evaluating the efficacy
of acupuncture, a convincing control and successful masking are critical.
Our study showed that adding a daily electroacupuncture procedure to
pharmacotherapy was more effective than pharmacotherapy alone in preventing
chemotherapy–induced emesis. Similar results recently have been observed
in an animal model.21 Are there biologically
plausible explanations for the observed effect? Pharmacological therapies
to manage chemotherapy–induced emesis have been directed at the neurotransmitter
receptors in the brain regions receptive to emetic stimuli.22,23
In a multiple-day, multiple-drug combination chemotherapy setting, emetic
responses are aggravated by the multiple agents, their converted products,
altered metabolism, and products of cell damage. The complex, multifactorial
and severe nature of such chemotherapy–induced emesis suggests that
no single antiemetic agent targeting a particular mechanism can be expected
to control vomiting completely. Electroacupuncture (repeated sensory stimulation)
has been thought to modulate serotonin, substance P, and endogenous opiates
along various pathways in the central nervous system.24- 27
We speculate that some of the effects we observed may be manifested through
the serotonin- and substance P–mediated components of the emetic reflex,
as well as through the opiate µ receptor via its antiemetic actions.28- 30 Future neurophysiological
and neurochemical investigations may help us to further understand the complexity
of emesis and to broaden the current approach to the spectrum of antiemetic