A total of 44 patients in the gabapentin group and 46 patients in the placebo group remained in the study for more than 1 night of admission and were included in the final data analysis.
Error bars represent SEMs. POD indicates postoperative day.
eAppendix. Enrollment Questionnaire
Customize your JAMA Network experience by selecting one or more topics from the list below.
Townsend M, Liou T, Kallogjeri D, et al. Effect of Perioperative Gabapentin Use on Postsurgical Pain in Patients Undergoing Head and Neck Mucosal Surgery: A Randomized Clinical Trial. JAMA Otolaryngol Head Neck Surg. 2018;144(11):959–966. doi:10.1001/jamaoto.2018.0282
Can perioperative pain control be improved with the addition of gabapentin to traditional narcotic medication regimens in patients undergoing mucosal head and neck surgery?
In this randomized clinical trial, compared with a placebo group of 46 individuals, 44 patients who received 300 mg twice daily of perioperative gabapentin had no difference in narcotic use but experienced less subjective pain.
Patients who undergo head and neck mucosal surgery and receive perioperative gabapentin treatment perceive less postoperative pain.
Effective postoperative pain management increases patient satisfaction, reduces cost, reduces morbidity, and shortens hospitalizations. Previous studies investigating multimodal pain therapy in otolaryngology patients focused on homogenous patient groups with short postoperative follow-up times.
To investigate the effect of perioperative gabapentin treatment on postsurgical pain in patients undergoing head and neck mucosal surgery.
Design, Setting, and Participants
Adults undergoing head and neck mucosal surgery from July 25, 2016, through June 19, 2017, were included in this double-blinded, placebo-controlled randomized clinical trial and randomized to receive gabapentin, 300 mg twice daily, or placebo before surgery and up to 72 hours after surgery.
Main Outcomes and Measures
Primary outcome was hourly narcotic use calculated in morphine equivalents. Secondary outcomes included subjective visual analog scale pain scores captured for resting, coughing, and swallowing using a 0- to 100-mm scale (a 100-mm line anchored with no pain on the left end and worst possible pain on the right end). A change of 10 mm or more was deemed to be clinically meaningful. Additional secondary outcome measures included degree of pain control, patient satisfaction, and adverse effects.
Of the 110 patients randomized to receive gabapentin or placebo, 11 and 10 withdrew from each group, respectively. Ninety patients were then analyzed: 44 in the gabapentin group (mean [SD] age, 61.1 [10.0] years; 33 [75%] male; 40 [91%] white) and 46 in the placebo group (mean [SD] age, 60.9 [11.3] years; 35 [78%] male; 43 [94%] white). Both groups had similar self-reported levels of preoperative pain and narcotic effectiveness. A median difference of 0.26 mg/h of morphine (95% CI, −0.27 to 0.94 mg/h) was found between groups. After controlling for comorbidity and self-reported baseline pain levels, mixed model analysis found the difference in marginal means of visual analog scale scores between groups to be lower in the gabapentin group compared with the placebo group for all categories (rest difference, 7.9 mm; 95% CI, −0.4 to 16.2 mm; cough difference, 8.9 mm; 95% CI, −0.5 to 18.3 mm; swallow difference, 9.4 mm; 95% CI, −1.2 to 20.0 mm). More patients in the gabapentin group reported that pain was always well controlled than in the placebo group (difference, 9.2%; 95% CI, −21% to 3%). Gabapentin and placebo groups reported similar levels of satisfaction with pain control (difference, 2%; 95% CI, −11% to 15%). There was no clinically meaningful difference in reported nausea between the 2 groups (difference, 6%; 95% CI, −14% to 26%).
Conclusion and Relevance
Perioperative gabapentin given 300 mg twice daily did not result in reduced narcotic use, but results were compatible with clinically meaningful reductions in pain scores. Satisfaction with pain control and adverse effects were similar between groups.
ClinicalTrials.gov Identifier: NCT02926573
Otolaryngology patients often undergo painful and functionally debilitating operations of the head and neck because of the neuroanatomy of this region. As a result, they are at increased risk for developing persistent postsurgical pain, dysphagia, and other complications.1,2Quiz Ref ID Effective management of acute postoperative pain reduces postoperative morbidity, increases patient satisfaction, shortens hospitalizations, and reduces hospital costs.1 Narcotic medications used to treat postsurgical pain are associated with constipation, nausea, and long-term addiction.3 Optimizing pain control to improve recovery and avoid overuse of narcotics is an important postoperative goal.
Quiz Ref IDMultimodal therapy, a widely accepted practice within pain management, uses nonnarcotic medications to improve pain and decrease narcotic requirements.4,5 Various surgical cohorts have demonstrated the benefits of multimodal therapy in acute postoperative pain.6-11 For instance, a randomized clinical trial by Rafiq et al6 found that administration of gabapentin, corticosteroids, ibuprofen, and paracetamol to postsurgical cardiac patients resulted in significant reductions in subjective pain scores. A review by Kehlet et al7 suggests that, in postsurgical patients, neural injury and subsequent central sensitization contribute significantly to chronic pain. Gabapentin, a medication that targets neuropathic pain, has been investigated as a postoperative pain adjunct.8,12-14Quiz Ref ID Meta-analyses of randomized clinical trials that evaluated gabapentin in a 300- to 1800–mg range, given as a single preoperative dose and as multiple perioperative doses, reported reduced odds of developing chronic pain (pooled odds ratio, 0.52; 95% CI, 0.27-0.98),13 reduced acute postoperative opioid consumption equivalent to 30 mg of morphine,9 a 35% reduction in total opioid consumption during the first 24 hours after surgery (ratio of means, 0.65; 95% CI, 0.59-0.72), and significantly reduced subjective pain on a 100-point visual analog scale (VAS).15,16
Randomized clinical trials that investigated gabapentin use in otolaryngology patients primarily used 1 or 2 preoperative doses that ranged from 600 to 1200 mg and a 24-hour postprocedure follow-up.17-22 Reductions in narcotic use were variable in effect size, ranging, for example, from 6 mg of morphine for 8 hours to 15 mg for 24 hours.18,22 Subjective pain scores varied from significantly reduced in the gabapentin group at all time points up to 24 hours,19,22 reduced with rest but not with motion up to 24 hours,17 and reduced only with motion and within 4 hours of surgery.20 Included surgical procedures were short, were conducted in the outpatient setting, and studied homogenous patient populations. Only 1 open-label trial investigated gabapentin use in patients undergoing large head and neck procedures.23
The goal of this study was to investigate gabapentin use in patients undergoing larger head and neck mucosal operations associated with significant risk of postoperative pain and dysphagia.24 A 300-mg dose of gabapentin given twice daily was selected and continued for 3 days including the day of surgery. We selected narcotic use as an objective and quantifiable primary outcome measure and used the VAS as a secondary outcome because of its ease of use and validity.
In this double-blinded, placebo-controlled randomized clinical trial, 250 adults were screened for participation, and 123 adults undergoing head and neck mucosal surgery from July 25, 2016, through June 19, 2017, were enrolled. Screening eliminated 74 patients because of inappropriate surgery, prior gabapentin use, or outpatient surgery status; 47 refused participation, and 6 had baseline chronic pain. A total of 123 patients were enrolled, 13 of whom withdrew (8 because of patient preference and 5 because of cancellation of surgery). Written informed consent was given in the outpatient clinics, and data collection was performed in the inpatient departments and surgical recovery areas. Patients 18 years or older with a glomerular filtration rate of greater than 30 mL/min/1.73 m2, no history of dementia, no history of chronic pain (defined as ≥6 months of pain or current pain unrelated to their index case), no current gabapentin use, and at least 1 night of inpatient admission planned were included. Patients were not excluded based on type of surgery or reconstruction if the primary indication involved a mucosal surface. Participants were randomly assigned to the gabapentin or the placebo group using a computer-generating scheme with blocks of varying sizes. All study personnel were masked to group assignment and intervention except the research statistician (D.K.) and pharmacist. Study protocols and consenting processes were approved by the Washington University School of Medicine in St Louis Institutional Review Board. Patients provided written informed consent, and all data were deidentified. The trial protocol can be found in Supplement 1.
Self-reported baseline pain and perceived opioid effectiveness were elicited from participants at enrollment using a short questionnaire asking participants to circle yes or no to the presence of daily pain aside from their head and neck condition and to rate narcotics as not at all effective to extremely effective (eAppendix in Supplement 2). Overall severity of comorbidity was captured using the Adult Comorbidity Evaluation-27.25 Participants received liquid gabapentin, 300 mg, or an equivalent volume of placebo twice daily per tube or orally. Days in the study were termed postoperative days (PODs) 0 (date of procedure) through 3 (last day of participation possible), and participants remained in the study up to POD 3 as long as they remained admitted and had no reason for study withdrawl. Participants received 1 dose of assigned intervention in the preoperative holding area 1 to 2 hours before surgery. Participants received their second dose of medication at 7 pm, if extubated by 5 pm. If the case went later than 5 pm, a second dose was not given that day. During PODs 1 to 2, participants received doses at 7 am and 7 pm in the inpatient department.
Subjective pain scores were captured using the VAS with resting, coughing, and swallowing motions.26 Strict scripting was used to gather scores and elicit adverse effects, and participants were asked to “Please rate your current pain level with no movement (rest), with a cough, and with a swallow.” Participants marked a point on a 100-mm line anchored with no pain on the left end and worst possible pain on the right end.26 A VAS score was recorded 1 to 2 hours after patient emergence from anesthesia and again at 7 pm on POD 0 if the emergence VAS was recorded before 7 pm. The VAS scores were recorded at 7 am, 10 am, and 7 pm on PODs 1 through 3 as long as the individual was still hospitalized and enrolled in the study. A discharge survey to document patient satisfaction and perceived pain control was given at the time of collection of the last VAS score.
Adverse effects associated with gabapentin, including dizziness, sedation, and nausea or vomiting, were elicited with each VAS score. Participants were followed up for adverse effects for 30 days after enrollment. Postoperative narcotic orders in the anesthesia recovery area and in the inpatient department were standardized for all patients. Acetaminophen, 1 g every 6 hours, was given to all participants. Participants received escalating doses of narcotic medications as clinically indicated.
Descriptive statistics were used to describe demographic, clinical, and treatment data between groups. Absolute and percentage differences, along with 95% CIs around the differences, were calculated to compare the gabapentin and placebo groups. Intention-to-treat analysis was performed, and safety analysis for adverse effects was undertaken after 50% enrollment was achieved and again at study completion. Masking was maintained for the safety analysis.
Primary outcome was the difference in mean per participant per group narcotic use per hour in oral morphine equivalents. Mean differences and 95% CIs were calculated for parametric data, and median differences and 95% CIs were calculated for nonparametric data. Secondary outcomes included VAS scores, patient satisfaction with pain control, and adverse effects. A hierarchical mixed model analysis with participant as the random factor was used to analyze VAS scores. This analysis makes maximum use of the data available at any time point without using a list-wise exclusion approach and allows for exploration of the time by group interaction both within and between participants. That is, this analysis explored changes in VAS scores between the 2 treatment groups across different time points.
We controlled for potential confounders of the model. Variables with a greater than 10% difference between groups were entered into the model and, if not significant, were excluded. The POD 0 VAS scores were excluded from analysis because of the high levels of sedation after anesthesia and the variable amount of narcotics given during surgery. A difference of 10 mm or greater in the VAS score is considered to be clinically meaningful.26 All statistical tests were 2-sided and evaluated at an α level of .05. SAS statistical software, version 9.4 (SAS Institute Inc) was used for all statistical analysis.
Sample size calculations were based on a study of the use of gabapentin in patients undergoing tonsillectomy.20 With use of relevant estimates from that study, it was determined that approximately 46 individuals per group (92 total) were sufficient to detect a difference of 20% or greater in mean morphine dose (effect size d = 0.6) between the placebo and gabapentin groups with an 80% power and a 2-sided α threshold value of .05.
Ninety patients completed the study, 44 in the gabapentin group (mean [SD] age, 61.1 [10.0] years; 33 [75%] male; 40 [91%] white) and 46 in the placebo group (mean [SD] age, 60.9 [11.3] years; 36 [78%] male; 43 [94%] white) (Figure 1). No significant differences were found in patient age, sex, race/ethnicity, comorbidity severity, tumor stage when applicable, reconstruction method, use of a skin graft, number of necks dissected, and self-reported narcotic effectiveness and daily pain among the individuals in the 2 intervention groups (Table 1). There was no difference between postoperative corticosteroid or aspirin use between the groups. Mean duration of anesthesia and time in the study were similar between the groups, and the number of individuals who received a second POD 0 medication dose was similar between the groups. Most individuals in each group remained in the study until POD 3 (Table 2).
Daily narcotic use in oral morphine equivalents was calculated for each participant for the total study duration. Quiz Ref IDThe median daily use for the entire study population was 1.58 mg/h (range, 0-6.11 mg/h). The median doses were 1.60 mg/h (range, 0-6.11 mg/h) in the gabapentin group and 1.59 mg/h (range, 0-5.84 mg/h) in the placebo group (median difference, 0.26; 95% CI, −0.27 to 0.94). When mean hourly doses were divided into the first POD (1.9 vs 1.8 mg/h), second POD (1.3 vs 1.6 mg/h), and when applicable, third POD (1.0 vs 2.0 mg/h), morphine use was again similar between the gabapentin and placebo groups. Corticosteroid use was similar between groups in the first, second, and third 24 hours after surgery and did not appear to be a confounding variable.
The mean VAS scores at each time point for gabapentin and placebo groups ranged from 22 to 46 mm with rest, 27 to 46 mm with cough, and 35 to 56 mm with swallow (Table 3). After controlling for differences in comorbidity and self-reported baseline pain levels, mixed model analysis revealed that marginal mean VAS scores in the gabapentin group were lower for all 3 categories (Figure 2). Quiz Ref IDThe marginal mean difference for the VAS rest score was 7.9 points lower (95% CI, −16.2 to 0.4 mm) in the gabapentin group compared with the placebo group. The CI reveals that the true difference could be as large as 16.2 mm lower. Likewise, the VAS cough score marginal mean difference was 8.9 mm lower (95% CI, −18.3 to 0.5 mm) in the gabapentin group compared with the placebo group, with a lower bound of the 95% CI showing that the true difference in the population could be as large as 18 mm lower. Finally, the VAS swallow score marginal mean difference was 9.4 mm lower (95% CI, −20.0 to 1.2 mm) in the gabapentin group compared with the placebo group, with the lower bound of the 95% CI showing that the true difference could be as large as 20 mm lower.
A total of 40 of the 42 participants (95%) in the gabapentin group reported that pain was always well controlled compared with 37 of the 43 participants (86%) in the placebo group (difference, 9.2%; 95% CI, −21% to 3%). Overall satisfaction with pain-related care was reported as very satisfied in 37 of 42 participants (88%) in the gabapentin group and 39 of 43 participants (90%) in the placebo group (difference, −2%; 95% CI, −15% to 11%). There was a slightly lower percentage of participants reporting sedation (difference, −5%; 95% CI, −15% to 25%) and a higher percentage reporting dizziness (difference, 5%; 95% CI, −9% to 19%) in the gabapentin group vs placebo group. A lower percentage of participants in the gabapentin group reported nausea compared with the placebo group (difference, 6%; 95% CI, −14% to 26%). No unanticipated adverse effects or harm was experienced by participants as a result of study participation.
Perioperative gabapentin administered at 300 mg twice daily for up to 72 hours was not associated with reduced narcotic use during the entire study period and when broken into 24-hour increments after surgery. This regimen resulted in improvements in subjective pain during rest, coughing, and swallowing. Similar levels of satisfaction with pain control were found between gabapentin and placebo groups, and a reduced incidence of nausea was seen in the gabapentin group.
Most otolaryngology studies17-22 that investigated gabapentin for acute postsurgical pain administered 600 to 1200 mg in a single preoperative dose and reported reduced pain medication use for up to 24 hours in patients undergoing tonsillectomy, thyroidectomy, and sinus surgery. Our cohort included more extensive operations, although postoperative pain levels were similar. Compared with our results, this finding suggests that a single, higher preoperative dose more effectively reduces narcotic use in these groups. The optimal regimen of gabapentin beyond 24 hours remains unclear. We chose a lower dose because we sought a regimen that could be stopped without a taper and that had a low risk of nausea or vomiting. Mikkelsen et al,27 in one of the few studies that administered gabapentin beyond 24 hours after surgery, found significantly higher rates of nausea or vomiting, sedation, and dizziness with 600 mg of gabapentin 3 times daily. We chose a lower dose of 300 mg twice daily for up to 3 days and did not find increased adverse effects compared with placebo. Future studies would investigate a higher dose during the 72-hour postoperative period, with an emphasis on monitoring adverse effects.
Improvements in subjective pain were seen in the gabapentin group in all 3 categories during all 3 postoperative days. A previous study28 found that postoperative pain is worse with movement, and we likewise found pain scores to be highest with swallowing. We found that gabapentin resulted in the largest reduction in pain when patients were asked to perform the most painful task. The VAS score changes between 11 and 15 mm are correlated with meaningful differences in pain levels, with a minimum change of 10 mm considered to be clinically meaningful in postsurgical patients in particular.26,29 In all 3 categories, reductions in pain were associated with effect sizes that reached clinically meaningful levels, and the estimated difference between means approached 10 mm in the swallow category. These reductions were seen at individual VAS time points and when combined differences in marginal means were compared in the mixed model analysis, which provides a point estimate of VAS score reduction that is applicable across the measurements. The larger improvements seen at later time points may represent a greater gabapentin effect with time. In addition, the patients remaining in the study up to POD 3 also underwent larger operations with extensive reconstructions and potentially are those who would be expected to have more postoperative pain.
Despite reductions in subjective pain in the gabapentin group, narcotic use remained similar between the groups. Studies that investigated the sensitivity of the VAS tool in postoperative patients potentially help elucidate this finding and report that scores in the 10- to 30-mm to 40-mm range correlate clinically with mild pain, in the 40- to 70-mm range with moderate pain, and in the greater than 70-mm range with severe pain.30-32 Changes in VAS scores in the mild range were found to be associated with escalating doses of narcotics in only 5% of cases.32 Changes to levels greater than 70 mm were associated with severe pain and correlated with increased doses 80% of the time. Levels in the 30- to 70-mm range were found to best correlate with narcotic increases when 10-mm or greater changes in scores occurred.32 Most pain scores in the gabapentin and placebo groups during all 3 categories remained in the mild to moderate range. Thus, our patients may have been overall less likely to ask for escalating narcotic doses, which made finding differences between groups more difficult to detect. In addition, we chose not to use patient-controlled analgesia to avoid the need to transition to an oral regimen, which could increase inpatient admission length. A patient-controlled analgesia potentially would have been better able to detect small differences in narcotic use rather than our oral regimen, which allowed for a range of doses.
Finally, we found similar rates of patient satisfaction between the groups. Anesthesia literature suggests that satisfaction with pain control is closely associated with the perception of attentive nursing as opposed to reported levels of pain.1 Our results are similar to other studies20,27 in otolaryngology that also found similar rates of satisfaction between gabapentin and placebo.
Limitations to our study include the multiple study personnel involved with the collection of VAS scores, which could have led to variation in data collection. To reduce this risk, strict scripting instructions were provided to all personnel. Several individuals had not previously taken narcotics and were unable to report baseline narcotic effectiveness in the preenrollment survey. Finally, the most common surgical sites were the oral cavity and oropharynx. Other head and neck mucosal sites are less well represented among the study participants; thus, the study results are not generalizable to patients who undergo surgery to other mucosal and nonmucosal head and neck sites.
Postoperative pain can limit function, slow recovery, and significantly affect quality of life. Superior methods for postsurgical pain control are an important area of investigation. We find that a low dose of perioperative gabapentin did not result in a reduction in narcotic use but was associated with reductions in subjective pain. This regimen has potential for patients who perceive significant postoperative pain and who are at risk for swallow dysfunction attributable to pain, but it is unlikely to change narcotic use.
Accepted for Publication: March 19, 2018.
Corresponding Author: Jay F. Piccirillo, MD, Department of Otolaryngology, Washington University School of Medicine in St Louis, 660 S Euclid Ave, McMillan Bldg, Campus Box 8115-06-805F, St Louis, MO 63110 (email@example.com).
Published Online: April 19, 2018. doi:10.1001/jamaoto.2018.0282
Author Contributions: Drs Townsend and Kallogjeri had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Townsend, Liou, Lindburg, Jackson, Bottros, Nussenbaum.
Acquisition, analysis, or interpretation of data: Townsend, Liou, Kallogjeri, Schoer, Scott-Wittenborn, Lindburg, Nussenbaum, Piccirillo.
Drafting of the manuscript: Townsend, Kallogjeri, Piccirillo.
Critical revision of the manuscript for important intellectual content: Townsend, Liou, Kallogjeri, Schoer, Scott-Wittenborn, Lindburg, Jackson, Bottros, Nussenbaum, Piccirillo.
Statistical analysis: Townsend, Kallogjeri.
Obtained funding: Townsend, Bottros.
Administrative, technical, or material support: Townsend, Liou, Scott-Wittenborn, Lindburg, Bottros, Nussenbaum, Piccirillo.
Study supervision: Jackson, Bottros, Nussenbaum, Piccirillo.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: Funding for this study was provided by grant 36435 from the Foundation for Barnes-Jewish Hospital, St Louis, Missouri (Dr Nussenbaum).
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: Dorina Kallogjeri, MD, MPH, is Statistics Editor and Jay F. Piccirillo, MD, is Editor of JAMA Otolaryngology–Head & Neck Surgery, but they were not involved in any of the decisions regarding review of the manuscript or its acceptance.
Meeting Presentation: This study was presented at the AHNS 2018 Annual Meeting; April 19, 2018; National Harbor, Maryland.
Additional Contributions: We acknowledge the research office coordinators with the Department of Radiology, Washington University School of Medicine in St Louis, St Louis, Missouri. Drs Townsend and Liou acknowledge the clinical research support from the Otolaryngology Surgical Outcomes and Quality Improvement Unit and the Clinical Outcomes Research Office at the Department of Otolaryngology–Head and Neck Surgery, Washington University School of Medicine in St Louis, St Louis, Missouri.