Preoperative Intradermal Acupuncture Reduces Postoperative Pain, Nausea and Vomiting, Analgesic Requirement, and Sympathoadrenal Responses

The protocol of this study was approved by the Institutional Review Board of University of Hirosaki (Hirosaki, Japan), and written informed consent was obtained from all participating patients. We studied patients scheduled for elective upper and lower abdominal surgery. We excluded those having one or more of the following conditions: (1) American Society of Anesthesiologists physical status greater than II; (2) preoperative opioid or epidural analgesia for cancer pain; (3) contraindication to epidural analgesia (i.e. , coagulopathy, skin infection); (4) body mass index greater than 30; (5) chronic obstructive or restrictive pulmonary disease; or (6) neurologic disease or immunocompromise.

We evaluated 107 patients undergoing elective upper abdominal surgery. On the day before surgery, an epidural catheter was inserted into the T7–T10 intravertebral space in the patients undergoing upper abdominal surgery. Epidural catheterization was performed with an 18-gauge Tuohy needle via  a median approach and the loss-of-resistance method. A 20-gauge catheter was inserted and advanced 5–7 cm; subarachnoid placement was excluded by injection of 3 ml of 2% lidocaine, and correct catheter position was tested by injection of 7–10 ml of 1% lidocaine. We next determined the area of analgesia by administering an additional 7 ml of 1% lidocaine. Patients were then observed for at least 90 min. Technical difficulties, analgesic dermatomes, and complications of the epidural technique were recorded.

After epidural catheterization, patients who chose to participate in the study were randomly assigned to acupuncture (n = 54) or control (n = 53). The randomized treatment assignment was maintained in sequentially numbered, sealed opaque envelopes. The envelopes were opened after epidural catheterization and confirmation of correct catheter placement.

A physician visited all patients 2 h before induction of anesthesia. Patients were positioned prone, and appropriate acupuncture sites for the planned surgery were identified with a skin marker. We chose 2.5 cm to the left and right of the T9–L3 spinal vertebrae (BL18–BL24). We thus chose a total of 14 points in each patient. Then we showed the 5-mm-long intradermal needles (Asahi Industry, Inc., Kawaguchi, Japan;fig. 1) to each patient in both groups. The diameter of these needles was 0.16 mm, and they had pencil-point tips. The loop at the end of each needle prevented it from completely entering the skin.

We also evaluated 84 patients undergoing lower abdominal surgery. Epidural catheterization was nearly the same as in those undergoing upper abdominal surgery, with an exception of choice of intervertebral space (T10–L1). After epidural catheterization, patients choosing to participate in the study were also randomly assigned to acupuncture (n = 41) and control (n = 41) groups. Before induction of anesthesia, intradermal needles were inserted with the same method used in patients undergoing upper abdominal surgery. We chose 2.5 cm to the left and right of the T11–L5 spinal vertebrae (BL20–BL26) in patients undergoing lower abdominal surgery.

We accurately explained to each patient that insertion of intradermal needles is virtually painless and that they may or may not feel slight pain during insertion. In patients undergoing both upper and lower abdominal surgery assigned to acupuncture, the intradermal needles were inserted almost horizontally into the skin at each acupoint. In patients in control groups for upper and lower abdominal surgery, a needle was positioned at each acupoint, but the needle was not inserted into the intradermal space. The needles were secured with opaque adhesive tape and remained in position for 4 postoperative days. Insertion of these thin, short intradermal needles is painless or very nearly painless. Because we inserted the needles to the back while patients were placed in prone position, they could not see our procedure. It is thus unlikely that individual patients were able to determine whether they were in the active treatment group. The physician in charge of intradermal acupuncture was, of course, aware of the group assignment. However, the anesthesiologists providing intraoperative care, physicians evaluating pain and opioid-related side effects, and the investigators measuring various biochemical mediators were fully blinded to the group assignments.

To assure that patients were blinded to needle insertion, we performed a preliminary study in 40 patients undergoing upper abdominal surgery who were randomly assigned to acupuncture (n = 20) or control (n = 20) groups. Intradermal needles were inserted in the acupuncture patients as previously described. On the third operative day, patients were asked which group they thought they were assigned to. In both groups, 50–60% of the patients were unable to guess to which group they had been assigned, 20–30% guessed acupuncture, and the remaining 20% patients thought they were in the control group.

Patients were premedicated with 0.01–0.02 mg/kg oral diazepam and 75 mg oral roxatidine (an H²blocker) 90 min before induction of anesthesia. Anesthesia was induced with 1–2 μg/kg fentanyl, 5 mg/kg thiopental, and 0.08 mg/kg vecuronium. We maintained anesthesia with 0.15 mg/kg droperidol, 10–20 μg/kg fentanyl, and vecuronium. During anesthesia, all patients were mechanically ventilated with 30% oxygen and 70% nitrous oxide to maintain the arterial carbon dioxide tension between 35 and 45 mmHg. Epidural anesthesia was not used intraoperatively. Radial arterial pressure, electrocardiogram, and pulse oximeter saturation were monitored in all patients. Lactated Ringer’s solution (8–10 ml · kg⁻¹· h⁻¹) was administered per the judgment of the attending anes-thesiologist.

We evaluated incisional pain during bed rest and during coughing, as well as deep visceral pain with a verbal rating scale (VRS; 0 = no pain, 1 = mild pain, 2 = moderate pain, 3 = severe pain). These pain parameters were finally graded to two categories: good pain relief (VRS = 0 or 1) or poor pain relief (VRS = 2 or 3). Patients assessed drowsiness, pruritus, and nausea and vomiting with a VRS (0 = none, 1 = mild, 2 = moderate, 3 = severe). One week after surgery, we asked patients to rate the overall adequacy of pain treatment: 0 = excellent, 1 = good, 2 = fair, and 3 = poor. We could not evaluate the incidence of urinary retention because each patient had an indwelling catheter for the initial postoperative days. Postoperative pain and analgesia-related complications (below) was evaluated in the recovery room before epidural morphine was administered (elapsed time zero), and at 8 am on the subsequent 4 postoperative days. After the first evaluation, we injected 3 mg morphine in 7–10 ml bupivacaine into the epidural catheter. Epidural morphine (3 mg) was then administered at 12-h intervals (9 pm and 9 am) for 3 postoperative days. No other epidural morphine was given.

Residual pain was treated with intravenous morphine as needed. We calculated the daily consumption of morphine during the first 4 postoperative days. The physician in charge of evaluating pain and analgesic-related complications visited the patients at approximately 8 am, 2 pm, and 8 pm for 4 postoperative days. When patients complained of pain at rest or during coughing (VRS ≥ 2), a bolus dose of 5 mg morphine was administered intravenously. If pain persisted (VRS ≥ 2), additional morphine was given at 30-min intervals until the VRS became less than 2. The maximum dose of morphine permitted between physician visits was 20 mg; the maximum daily allowance for intravenous morphine was thus 60 mg. On patients’ request, midazolam was given intravenously at 10 pm. Respiratory depression (arterial carbon dioxide tension > 55 mmHg) accompanied by drowsiness was treated with intravenous naloxone. Pruritus was treated with antihistamines. Metoclopramide was administered as necessary for treatment of nausea and vomiting.

Arterial blood for determination of cortisol, epinephrine, norepinephrine, and dopamine concentrations was sampled before induction of anesthesia, immediately before surgery, 1 h after the beginning of surgery, on emergence from anesthesia in the recovery room, and on the first postoperative day. The blood was anticoagulated with 10 mg EDTA and centrifuged immediately; the resulting plasma was stored at −80°C until analysis.

The plasma concentration of all hormones were measured in duplicate. Cortisol was measured in 0.5 ml of plasma by a slight modification of a spectrophotometric method. 11The excitation and emission wavelengths were 470 and 530 nm, respectively. The coefficient of variation was less than 5%. Catecholamines were measured by our established method. 12Briefly, catecholamines in 1 ml of plasma were extracted to activated almina (Wako, Tokyo, Japan) and concentrated in 150 μl methylalcohol containing 5% acetate. Pentafluoropropionic anhydride derivatives of catecholamines were synthesized and analyzed by gas chromatography–mass spectrometry (JMS-D300, JEOL, Tokyo, Japan) with a multiion detector. We used dopamine-D3 as an internal standard. A glass column contained 10% silicone GE SE-30 (80–100 mesh; GL Sciences Inc., Tokyo, Japan). The temperature of the column was 190°C. We used helium as the carrier gas at 1.2 kg/cm². The mass spectrum showed molecular ions at mass numbers 428, 431, 590, and 604, corresponding to the pentafluoropropionic anhydride derivatives of dopamine, dopamine-D3, norepinephrine, and epinephrine, respectively. Minimum detection concentrations of epinephrine, norepinephrine, and dopamine were 2, 1, and 1 pg/ml, respectively. The coefficient of variation was less than 7% for each.

The studies in patients having upper and lower abdominal surgery were considered to be independent because we could hardly randomly assign the type of surgery. We therefore avoided statistical comparisons between the types of surgery, but instead restricted analysis to the randomized treatment groups within each type of surgery. The first evaluation of pain in the recovery room was designated as elapsed time zero. Each pain parameter was evaluated with a chi-square test.

Patients in each group were allocated to one of two categories: those with good pain relief (VRS = 0 or 1) and those with poor pain relief (VRS = 2 or 3). Other time-dependent intragroup data were evaluated with repeated-measures analysis of variance and Dunnett tests for comparison to elapsed time zero;P < 0.05 was considered statistically significant. Differences between the control and acupuncture groups at each time point were evaluated with two-tailed, unpaired t  tests or chi-square tests, as appropriate. We compared values in the two groups at five or four (daily consumption of morphine) time points and thus used a Bonferroni correction. Data are expressed as mean ± SD.

We enrolled and randomized 107 patients for upper abdominal surgery and 82 for lower abdominal surgery. Fourteen patients were unavoidably excluded because postoperative artificial respiration was required to treat dyspnea, low partial pressure of oxygen, pulmonary embolism, myocardial ischemia, or massive bleeding. Therefore, final evaluation was completed on 98 patients for upper abdominal surgery and 77 for lower abdominal surgery. Demographic and morphometric characteristics and intraoperative management were similar in the acupuncture and control groups (tables 1 and 2).

More than 30% of patients received calcium blockers for intraoperative hypertension. However, we did not administer β blockers, clonidine, and anti–angiotensin-converting enzymes. No patients complained of pain during insertion of the intradermal needles. All evaluated patients emerged from anesthesia within 10 min after nitrous oxide was discontinued and muscle relaxation was antagonized. No complications such as skin infection or postinsertion discomfort developed in any patients during the 5-day study period. One week after surgery, more than 95% of patients in both groups gave an excellent-to-good judgment of our overall pain treatment.

The degree of pain relief on the VRS for incisional pain at rest and during cough and for deep visceral pain improved significantly over time in patients undergoing upper abdominal surgery in both groups (P < 0.0001). However, from the recovery room to the second postoperative day, postoperative pain relief was significantly better in the acupuncture than in the control group (P < 0.05;table 3). Results were nearly similar after lower abdominal surgery: analgesia was significantly better in the acupuncture than in the control groups from the recovery room to the first or second postoperative day (P < 0.05;table 4). In patients undergoing both upper and lower abdominal surgery, daily consumption of morphine decreased over time in both groups (P < 0.0001) but was up to 50% less in the acupuncture than in the control group on postoperative days 1–4 (P < 0.01;fig. 2).

None of the patients experienced respiratory depression. The incidence of other analgesic-related side effects was greatest on the day of surgery and subsequently decreased gradually. None of the observed side effects was severe. The incidence of drowsiness and pruritus was similar in the acupuncture and control groups. Postoperative nausea and vomiting were less frequent after upper abdominal surgery in the acupuncture than control groups (24%vs.  50%;P < 0.05). The incidence of nausea and vomiting was also less in the acupuncture than control patients after lower abdominal surgery (18%vs.  47%;P < 0.01;fig. 3).

Starting 1 h after the beginning of upper abdominal surgery, plasma concentrations of cortisol, norepinephrine, and epinephrine increased significantly in both groups. Plasma dopamine concentrations were elevated during recovery in both groups. All catecholamine concentrations except norepinephrine peaked at emergence from anesthesia. The increases in plasma concentrations of epinephrine and cortisol were greater in the control than in the acupuncture group during recovery and the subsequent day (P < 0.01;fig. 4). The results were comparable during lower abdominal surgery. However, plasma dopamine concentrations failed to increase in either group (fig. 5).

Acupuncture increased the fraction of patients with good pain relief from the recovery room to the second postoperative day and reduced analgesic requirements by up to 50%. The mechanism by which acupuncture produces analgesia remains unclear. However, previous clinical studies 1–4are consistent with a classical “gate control theory” in which sensory stimulation by intradermal needles activates large nerve fibers and changes pain perception in the spinal cord. This then diminishes pain stimulation transmitted via  smaller fibers. Secretion of endogenous opioids such as endorphins, enkephalins, and dynorphins may also contribute to acupuncture analgesia. 13 

Our study improves on previous ones 1–5because we performed intradermal acupuncture before anesthesia and surgery and thus did not interfere with operative maneuvers and postoperative rest. Our needles were so small and so fine that none of the patients complained of needling pain or postneedling soreness. Intradermal acupuncture is easy to use and requires no special training. Furthermore, the cost/performance ratio is better than with electric acupuncture because a stimulator is not required. An important aspect of our study is that it was double-blinded. Consequently, neither the patients nor the evaluating physicians were aware of the treatment status. This is a key element of the study design because placebo effects in acupuncture can be substantial and have confounded numerous previous investigations. 13,14 

There are two types of pain after abdominal surgery: incisional and visceral. In our experience, sharp incisional pain dominates the early postoperative course. Deep visceral pain thus appeared to be masked by the sharp incisional pain on the first operative day. Incisional pain was successfully treated by opioid analgesics, but deep visceral pain was not relieved and thus dominated after several days. Although acupuncture may have modulated the visceral nociceptive system in animal experiments, 15,16whether acupuncture improves postoperative visceral pain has not been clear. One study by Smith et al.  17suggested that electroacupuncture was more effective in reducing incisional pain than in relieving visceral pain. An important finding in our study is that preoperative acupuncture reduced deep visceral pain. This is consistent with the fact that stimulation of acupoints of BL is traditionally used for various types of pain caused by vital organ disturbances.

Preoperative intradermal acupuncture reduced the incidence of postoperative nausea and vomiting after upper and lower abdominal surgery. This antiemetic effect can be explained, in part, by better analgesia and a lesser opioid requirement. 1,2,18In addition to analgesic effects, direct antiemetic effects are also plausible. Dundee et al.  19reported that acupuncture needling of the sixth point on the pericardial meridian (P6, Neikuan) is antiemetic in adult patients undergoing minor gynecologic surgery in whom pain was unlikely to be a major trigger for nausea and vomiting. Ho et al.  18reported the beneficial effect of acupuncture on prevention of nausea and vomiting induced by epidural morphine.

In contrast, two studies 20,21showed that acupuncture to P6 was not beneficial for postoperative nausea and vomiting after strabismus surgery where pain is not the trigger. This inconsistency can be explained, in part, by timing of acupuncture and selection of acupoints. Dundee et al.  19pointed out that to be effective, acupuncture should be performed before emetic stimulation by anesthesia and surgery. Veroli and Astier 22pointed out the importance of the choice of acupoints: bladder and gall bladder points are better than P6 for treatment of nausea and vomiting. The marked antiemetic effect we observed is thus presumably in part a result of our selection of acupoints and preanesthetic needling.

Thoracic epidural anesthesia can completely block the endocrine response to lower abdominal surgery. 9,10However, high thoracic epidural anesthesia failed to suppress the responses. 9,10Segawa et al.  23showed the importance of afferent stimulation of the phrenic nerve for endocrine responses in upper abdominal surgery. Few studies evaluate the effects of acupuncture on the adrenal hormones. 24,25Plasma cortisol and epinephrine concentrations increased over time and peaked in the recovery room, but the increases were smaller in patients who received acupuncture than in those who did not. Postoperative pain increases the plasma concentrations of cortisol and epinephrine. 26–29It is thus likely that cortisol and epinephrine concentrations increased less in the treated than untreated patients because acupuncture significantly reduced postoperative pain.

Intraoperative increases in plasma norepinephrine were similar in the acupuncture and control groups. Plasma norepinephrine does not correlate as closely with pain as do epinephrine and cortisol. Camu and Debucquoy 30reported that plasma concentrations of epinephrine and cortisol decreased after postoperative opioid administration, whereas norepinephrine did not change. According to Tschernko et al. , 6postoperative pain intensity and the analgesic requirement were much less in patients undergoing thoracoscopic surgery than in those who underwent axillary thoracotomy. Significantly greater plasma epinephrine concentrations were noted in the former than in the latter patients, but there were no significant differences in plasma norepinephrine concentrations. Plasma dopamine concentrations were also similar in the acupuncture and control groups. Plasma dopamine concentrations increase when the adrenal medulla is maximally stimulated. 31Our results suggest that acupuncture cannot suppress a marked adrenomedullary activation by intraoperative nociceptive stimulation.

Although acupuncture per se  reduces the sympathoadrenal activity in both animal 32and human 33,34studies, our results differ from those reported by Kho et al. , 25who reported no differences in plasma adrenal hormones in patients with and without intraoperative acupuncture. This inconsistency may be a result, in part, of methodologic differences. Moreover, a serious limitation of their study is that they failed to evaluate postoperative pain or analgesic requirements. It therefore remains unclear whether their acupuncture method produced postoperative analgesia.

In summary, preoperative intradermal acupuncture does not require special training, does not interfere with operative maneuvers or postoperative rest, and does not cause complications. Intradermal acupuncture decreased incisional and visceral pain, reduced analgesic requirement, and diminished the incidence of postoperative nausea and vomiting. Acupuncture analgesia also reduced the pain-induced activation of the sympathetic nervous system that normally accompanies surgery. We conclude that preoperative intradermal acupuncture is easy to use, safe, and markedly improves postoperative analgesia.

The authors thank Elisabeth F. Lanzl, A.B., A.M. (University of Chicago, Chicago, Illinois), for editing the manuscript.