Gabapentin for Prophylaxis Intrathecal Morphine-Induced Pruritus

The study was approved by the ethic committee of Songklanagarind Hospital in Songkhla, Thailand, and was designed as a randomized, double-blind, placebo-controlled trial. All patients gave their informed consent after receiving written information of the study. Orthopedic patients were invited to participated between September 2009 and August 2010. Patients were considered eligible if they were aged between 15-70 yr, ASA physical status I-III, and were scheduled for lower limb surgery under spinal anesthesia. Patients were excluded if they presented of any following reasons: contraindication for spinal anesthesia, known allergy history to gabapentin, complaint of pruritus before surgery, morbid obesity (BMI > 35), coexisting skin disorder, and any systemic disease associated with pruritus. Patients who had history of seizure attacks, mental illness, chronic headache, or neuropathic pain and were concomitantly using of anticonvulsants, antidepressants, antipsychotics, or antihistamine were also excluded. Patients were randomized into the treatment and the placebo groups according to a computer-generated randomization list and patients were allocated consecutively. The blinding was maintained by keeping the treatment code with one of the investigators, separating them from the investigator performing the assessments. Patients in the gabapentin group received two capsules of gabapentin 300 mg (Neurontin®, Pfizer) at 2 h before operation. The patients in the placebo group received equal numbers of identical looking placebo, according to the same schedule. In order to prepare the identical looking medications, for gabapentin group we coated gabapentin capsules by other capsules and for placebo group we filled the identical looking capsules with flour. All patients did not receive any other sedative agents beside the study drugs. After standard monitoring (electrocardiogram, noninvasive arterial blood pressure, and pulse oximetry) was set-up in the operating room, each patient was received Ringer's lactate or normal saline solution 5-10 mL/kg, spinal anesthesia was performed at the L2-3 or L3-4 interspace with a 27-gauge Quincke-type needle using 0.5% isobaric bupivacaine plus 0.2 mg of preservative-free morphine. Midazolam, morphine, fentanyl and propofol were administered intravenously (IV) for intraoperative sedation at the discretion of the anesthesiologist. We excluded the patient who failed spinal block, inadequate block, and the patient who had prolong operative time which had to continue with general anesthesia because we could not evaluate pruritus. The sedation level was evaluated by the Ramsay Sedation Scale during the operation. The patients were followed for 24 h after intrathecal administration of morphine. Postoperative wound pain was assessed with a verbal numeric rating scale (VNRS). Rescue treatments for postoperative pain were provided with any medications ordered by anesthesiologists or orthopedists. Pruritus was evaluated at 1, 2, 3, 4, 6, 9, 12, and 24 h after intrathecal administration of morphine by a blinded investigator. The patients were questioned about the presence and degree of pruritus. The degree of pruritus was classified as 0 = absent, 1 = mild (restricted to one area such as face or arms, not troubling the patient, often reported only after prompting), 2 = moderate (affecting a larger area such as face and arms or face and anterior surface of thorax, not disturbing the patient, therefore not requiring treatment), or 3 = severe (extensive or generalized, often disturbing the patient to the point of necessitating treatment). Severe pruritus was treated with 10 mg IV chlopheniramine. Patients were also evaluated the severity of postoperative nausea and vomiting (PONV) which graded on a four point scale as 0 = no nausea or vomiting, 1 = mild nausea only, 2 = nausea or vomiting responding to initial treatment, and 3 = nausea or vomiting requiring repeat treatment, the presence of urinary retention and the other side effects of drug treatment. Patients who reported vomiting received 4 mg IV ondansetron or 10 mg IV metoclopramide. The primary outcome measure of the study was the incidence of pruritus during the 24 h follow-up period. The difference of onset time and severity of pruritus in the gabapentin and placebo groups served as the secondary outcome measure. Additional secondary outcome measures were side effects of the gabapentin group. There was an intention-to-treat in our study. Statistical analysis was performed using the R 2.11.1 software. We considered a 35% reduction in the incidence of pruritus to be clinically important according to the previous study's result showed that gabapentin 1200 mg could reduce incidence of intrathecal morphine-induce pruritus by 38%12. In our study we reduced the dose of gabapentin to 600 mg as describe previously so we think the efficacy of the drug may decrease so we assumed that gabapentin 600 mg could reduced incidence of intrathecal morphine-induced pruritus by 35%. Power analysis was performed to determine the sample size with a probability for a type II error of 0.2 and type I error of 0.05. To detect a 35% reduction in the incidence of pruritus, using the incidence of pruritus in Songklanagarind Hospital which was 63%, a sample size of 80 patients in each group was estimated to be required. To accommodate for 10% patient dropouts and failure of spinal anesthesia, 88 patients were enrolled in each group. Incidence of pruritus was measured by using Chi-Square test. The onset time and severity of pruritus was analyzed by means of Kaplan-Meier probability curves. Continuous data were analyzed using un-paired student T-test for normal distribution data such as BMI and by Mann-Whitney U test for non normal distribution data such as age, operative time, postoperative verbal numeric rating pain score. Comparison of the categorical data was performed using chi-square test such as incidence of PONV, hypotension, and shivering and by Fisher's exact test if expected value < 5 such as type of operation, incidence of bradycardia, and ASA classification. Results are expressed as mean+SD for normal distribution data and median (range) for non normal distribution. For all determinations, P values of < 0.05 were considered to be statistically significant.