Introduction
Transoral robotic surgery offers patients with human papilloma virus (HPV) associated oropharyngeal squamous cell carcinoma (SCC) good long-term prognosis and is becoming increasingly utilised in place of traditional surgical techniques. However, immediate post-operative dysphagia occurs frequently, which can prolong hospital stay or result in re-hospitalisation secondary to dehydration from inadequate oral intake, and can also lead to weight loss.Reference Achim, Bolognone, Palmer, Graville, Light and Li1 Frequently, nasogastric tubes are placed during surgery to mitigate the effects of dysphagia and allow patients to maintain adequate oral intake.Reference Achim, Bolognone, Palmer, Graville, Light and Li1 However, nasogastric tubes are uncomfortableReference Corry, Poon, McPhee, Milner, Cruickshank and Porceddu2 and increase the complexity of hospital discharge.
While there are several mechanisms contributing to post-operative dysphagia following transoral robotic surgery, pain (and resultant anxiety) is a major contributing factor.Reference Albergotti, Jordan, Anthony, Abberbock, Wasserman-Wincko and Kim3 Typically, opioid analgesics are used to control post-operative pain following transoral robotic surgery. However, opioids have many undesirable side effects.Reference Trasolini, McKnight and Dorr4 In addition, their use can lead to dependency in up to 6 per cent of patients who receive an opioid for an acute care episode,Reference Chen, Hedegaard and Warner5 and can lead to death in the community.Reference Gomes, Greaves, van den Brink, Antoniou, Mamdani and Paterson6 These concerns have raised interest in post-operative multimodal, opioid-sparing analgesia regimens.Reference Trasolini, McKnight and Dorr4 Our traditional practice relied on opioid medications and scheduled acetaminophen to provide post-operative analgesia following transoral robotic surgery for oropharyngeal SCC. In order to decrease opioid administration, we introduced scheduled post-operative gabapentin and ‘as needed’ ibuprofen to our analgesia regimen.
Gabapentin, an anti-convulsant with anti-nociceptive properties, provides modest post-operative analgesia and opioid-sparing effects.Reference Bialer7–Reference Mishriky, Waldron and Habib10 Particularly germane to transoral robotic surgery, Starmer et al. demonstrated that when gabapentin was administered to oropharyngeal SCC patients undergoing chemoradiation, they began using their percutaneous endoscopic gastrostomy (PEG) feeding tubes later, had their PEG tubes removed earlier and had improved objective swallowing scores.Reference Starmer, Yang, Raval, Gourin, Richardson and Kumar11 Non-steroidal anti-inflammatory drugs (NSAIDs) such as ibuprofen also have post-operative opioid-sparing properties.Reference Trasolini, McKnight and Dorr4 We traditionally avoided ibuprofen because of concern for post-operative haemorrhage,Reference Pollei, Hinni, Moore, Hayden, Olsen and Casler12 but recent evidence has documented the safety of NSAIDs following tonsillectomy.Reference Bedwell, Pierce, Levy and Shah13–Reference Pfaff, Hsu and Chennupati16
Our anecdotal impression was that gabapentin and ibuprofen provided better post-operative pain control than opioid analgesics following transoral robotic surgery. This retrospective study reviewed the post-operative analgesia course of patients undergoing transoral robotic surgery for oropharyngeal SCC to test the hypothesis that these medications were associated with improved post-operative pain relief.
Materials and methods
This study was approved by the Mayo Clinic Institutional Review Board (approval number: 14-004693). In line with Minnesota Statute 144.295, all patients in this study provided prior authorisation to use their medical records for research.
Study design and patient selection
The data in this study represent a high-volume transoral robotic surgery practice at a tertiary academic institution. We performed a retrospective review of our departmental oropharyngeal SCC RedCap (Research Electronic Data Capture) database to identify adult patients (aged 18 years or older) who underwent transoral robotic surgery for a primary oropharyngeal SCC from 5 September 2012 to 30 November 2016. Surgical care was provided by a group of six otolaryngology – head and neck surgeons.
Because the introduction of gabapentin represented the initial change in our multimodal regimen, and the introduction of ibuprofen occurred over a slower transition period, we identified the initial date of gabapentin introduction as March 2015. The study included all patients treated over a four-year period to represent a contemporary group of patients treated prior to (late 2012–2015) and following (2015–2016) the introduction of gabapentin into our practice. Patients were excluded if they had a history of prior head and neck cancer.
During the study period, there were no meaningful changes in the surgical techniques used for transoral robotic surgery oropharyngectomy or extent of neck dissection.Reference Moore, Olsen and Kasperbauer17,Reference Moore, Olsen and Martin18 However, there was a shift towards an earlier initiation of post-operative swallowing, with the goals of limiting nasogastric tube duration and reducing the number of patients discharged from hospital with nasogastric tubes.
Multimodal analgesic protocols
Peri-operative analgesia regimens were dependent on provider preference. Post-operative analgesic medications included: acetaminophen (1000 mg scheduled every 6 hours), intravenous (IV) opioids (typically for the first post-operative 12 hours), oral oxycodone (typically available as 5–10 mg every 2–6 hours as needed for pain), gabapentin (initiated on post-operative day 1 and titrated from 300 mg in divided doses on post-operative day 1, to 600 mg on post-operative day 2, and 900 mg on post-operative day 3), and ibuprofen (200–600 mg as needed every 6 hours).
In our practice, most patients are discharged home by post-operative day 3; therefore, analysis focused on post-operative days 1 and 2, to decrease the bias of intra-operative medications (post-operative day 0) and day of discharge.
Post-operative care
Post-operative transoral robotic surgery patients are admitted to a general post-operative ward staffed with otolaryngology – head and neck surgery trained registered nurses. Vital signs are assessed on admission, 30 minutes following arrival, then hourly for 4 consecutive hours, then every 4 hours for 48 hours, and then every 8 hours until discharge. Pain scores are recorded on a numerical pain scale of 0 (no pain) to 10 (worst pain imaginable). Sedation is assessed with the Richmond Agitation and Sedation Scale.Reference Sessler, Gosnell, Grap, Brophy, O'Neal and Keane19 Pain, sedation level and vital signs are checked prior to IV opioid administration, and then 5 minutes and 30 minutes after administration. In patients receiving opioids, nurses will apply oxygen to keep oxygen saturation greater than 90 per cent, and administer naloxone if needed. If patients have a respiratory rate of less than 8 per minute or a Richmond Agitation and Sedation Scale score of −3 or lower, the nurses contact the otolaryngology – head and neck surgeon team.
Nasogastric tubes are placed at the discretion of the surgeon based on the extent of the surgery and the co-morbid status of the patient. Patients are routinely started on an oral diet on post-operative day 0 or 1, advancing from liquid to mechanical soft. Once patients can take oral medications and their maintenance fluids by mouth (typically 600 ml), the nasogastric tube is removed. If patients are not able to meet these criteria, tube feeds are initiated and continued until oral intake improves as an in-patient, or the patient is cleared by speech pathology to return to an oral diet as an out-patient; this typically involves a modified barium video swallow assessment at one to two weeks following surgery. Management of an intra-operative pharyngotomy has been previously described.Reference Moore, Olsen and Martin18
Data abstraction
Electronic medical records were abstracted using previously described proprietary software,Reference Herasevich, Kor, Li and Pickering20 as well as from our otolaryngology – head and neck surgery oropharyngeal SCC RedCap database. Pre-operative patient variables included age, sex, smoking status, adult co-morbidity evaluation (using the Adult Comorbidity Evaluation-27 index)Reference Piccirillo, Tierney, Costas, Grove and Spitznagel21 and home analgesic use. Surgical variables included duration of surgery, cumulative dose of intra-operative opioids (reported as oral morphine equivalents),22 laterality of neck dissection and presence of an intra-operative pharyngotomy. Oncological variables included HPV status (positive p16 and/or HPV in situ hybridisation), primary tumour subsite, pathological size of the primary tumour and tumour (T) stage (based on the eighth edition of the American Joint Committee on Cancer guidelinesReference Lydiatt, Patel, O'Sullivan, Brandwein, Ridge and Migliacci23). Post-operative data included hospital length of stay, analgesics, severe pain (defined as a pain score of 7 or higher), sedation (defined as Richmond Agitation and Sedation Scale score of –1 or lower), the occurrence of complications, and hospital re-admission within 30 days of surgery.
Statistical analysis
Data are presented as mean (standard deviation) or median (interquartile range) values for continuous variables and number (percentage) values for categorical variables. The primary end points were the dose of post-operative opioids and binary variables indicating severe pain and sedation on post-operative days 1 and 2. We compared patients with and without these events using the student's t-test or rank sum for continuous variables and Fisher's exact test for categorical variables. In addition, multivariable logistic regression analyses were performed to assess potential associations between abstracted variables and post-operative opioid dose, severe pain and sedation. The results of the multivariable model are summarised with effect size estimates for opioid dose, and odds ratios for severe pain and sedation, with 95 per cent confidence intervals. Statistical analysis was performed using JMP Pro software, version 14.1.0 (SAS Institute, Cary, North Carolina, USA).
Results
A total of 216 patients underwent transoral robotic surgery procedures and were included. Table 1 presents demographic, oncological and surgical characteristics.
Table 1. Demographic, oncological and surgical characteristics of oropharyngeal SCC patients who underwent TORS and neck dissection
*Patients reported giving up smoking more than one month prior to surgery. †Other = posterior oropharyngeal wall and soft palate. ‡One patient did not undergo neck dissection because of substantial co-morbidities and a clinically staged node-negative neck. SCC = squamous cell carcinoma; TORS = transoral robotic surgery; SD = standard deviation; HPV = human papilloma virus; ACE-27 = Adult Comorbidity Evaluation-27 index; NSAIDs = non-steroidal anti-inflammatory drugs; IQR = interquartile range
A total of 129 patients had a nasogastric tube placed during their hospital stay (typically during surgery), which was subsequently removed a median (interquartile range) of 3 (1, 15) days following surgery. A substantial reduction of nasogastric tube duration was noted after 2014 (median (interquartile range) of 8 (2, 18) vs 2 (1, 13) days, p = 0.003) and in those patients given gabapentin (2 (1, 3) vs 7 (2, 17) days, p < 0.001). Ibuprofen was not associated with reduced nasogastric tube duration.
Fifty patients (23.1 per cent) were discharged with a nasogastric tube. Patients undergoing more extensive surgery, as indicated by operative time, intra-operative opioid administration, pharyngotomy and bilateral neck dissection, were more likely to go home with a nasogastric tube. Patients discharged with a nasogastric tube had longer hospital stays and higher rates of anti-emetic administration.
Table 2 summarises post-operative analgesia administration, and the occurrence of severe post-operative pain or sedation. The number of patients experiencing severe pain was highest on post-operative day 0, and sedation was highest on post-operative day 2.
Table 2. Post-operative analgesia, severe pain and sedation in patients who underwent TORS for oropharyngeal SCC
*Decreasing number of patients reflects patients being discharged home, beginning on post-operative day 1. †Severe pain reflects a numeric rating scale score for pain of 7 or higher. ‡Sedation indicates a Richmond Agitation and Sedation Scale score of less than −1. TORS = transoral robotic surgery; SCC = squamous cell carcinoma; post-op = post-operative; IQR = interquartile range
Table 3 presents the associated estimated change in post-operative opioid use and clinical characteristics. A reduction of post-operative opioid requirements was associated with advancing age. An increase in post-operative opioid requirements was associated with greater intra-operative opioid administration and post-operative ibuprofen administration. Other patient, tumour and surgical variables, and post-operative gabapentin use, were not associated with post-operative opioid requirements.
Table 3. Analysis of oral morphine equivalent use on post-operative days 1 and 2 for patients who underwent TORS for oropharyngeal SCC
Data represent effect size estimate (95 per cent confidence interval) values, unless indicated otherwise. *Variables in increments of 10. †All other patients had ipsilateral neck dissection except one patient. ‡Corresponds to analgesics given post-operatively. TORS = transoral robotic surgery; SCC = squamous cell carcinoma; post-op = post-operative; ACE-27 = Adult Comorbidity Evaluation-27 index
Tables 4 and 5 summarise clinical differences in patients who experienced severe pain or sedation on post-operative days 1 and 2.
Table 4. Analysis of severe pain scores for 214 patients on post-operative days 1 and 2 following TORS for oropharyngeal SCC
*Severe pain reflects a numeric rating scale score for pain of 7 or higher. †Others = soft palate (n = 1) and posterior oropharyngeal wall (n = 1). ‡All other patients had ipsilateral neck dissection except one patient. TORS = transoral robotic surgery; SCC = squamous cell carcinoma; SD = standard deviation; ACE-27 = Adult Comorbidity Evaluation-27 index; HPV = human papilloma virus; IQR = interquartile range
Table 5. Analysis of sedation scores for 214 patients on post-operative days 1 and 2 following TORS for oropharyngeal SCC
*Sedation indicates a Richmond Agitation and Sedation Scale score of less than −1. †Others = soft palate (n = 1) and posterior oropharyngeal wall (n = 1). ‡All other patients had ipsilateral neck dissection except one patient. TORS = transoral robotic surgery; SCC = squamous cell carcinoma; SD = standard deviation; ACE-27 = Adult Comorbidity Evaluation-27 index; HPV = human papilloma virus; IQR = interquartile range
There were 12 cases (5.6 per cent) of post-operative haemorrhage, 2 occurred during hospitalisation and 10 following hospital discharge. Three of these cases were managed in the operating theatre; the remainder were managed with observation. Six of the patients who had post-operative haemorrhage received ibuprofen, but its use was not associated with bleeding rate (p = 0.221).
Following surgery, five patients were admitted to the intensive care unit (two from the post-operative care unit for haemodynamic instability, two from the ward for hypertensive urgency, and one for respiratory failure on post-operative day 2). The patient who developed respiratory failure was a 67-year-old male with a tonsil primary, tumour stage T4, requiring bilateral neck dissections. This patient developed pneumonia on post-operative day 2, and required tracheal intubation, mechanical ventilation and ultimately tracheostomy. He had been administered gabapentin post-operatively (post-operative day 0 = 0 mg; post-operative days 1–5 = 300 mg daily); his modified Richmond Agitation and Sedation Scale score on post-operative days 0–5 was 0, indicating no sedation.
Following hospital discharge, 26 patients required evaluation in the emergency department: 10 for bleeding (mentioned above), 6 for pain control, 3 for dehydration, and 1 each for wound dehiscence, tracheocutaneous fistula, feeding tube dislodgement, pulmonary embolism, neck swelling and anxiety (some patients had multiple reasons). Six additional patients were hospitalised (five for pain control and dehydration, and one to the intensive care unit for pulmonary embolism). There was no 30-day mortality.
Discussion
The primary finding in our study was that scheduled gabapentin was not associated with decreased rates of severe pain or post-operative opioid analgesic use. Another important observation was that ibuprofen was not associated with an increased haemorrhage rate but was administered less often in patients with severe pain, suggesting that it was under-utilised. Despite the potential benefits of post-operative multimodal analgesia for transoral robotic surgery patients, current literature does not address optimal analgesia regimens. In order to maximise pain control while minimising over-sedation in patients following transoral robotic surgery, post-operative analgesia must consider the patient's age, co-morbidities, and the side effects of both gabapentin and ibuprofen.
The effectiveness and safety of gabapentin following transoral robotic surgery are unknown. In a meta-regression analysis, peri-operative gabapentin demonstrated a degree of opioid-sparing effect of varying clinical importance for different surgical procedures.Reference Doleman, Heinink, Read, Faleiro, Lund and Williams8 Mishriky et al. reviewed 55 studies, and found that pregabalin improved post-operative analgesia but also increased sedation.Reference Mishriky, Waldron and Habib10 Concerns regarding gabapentinoid-associated sedation have been heightened, especially in regard to respiratory depression with concomitant opioid administration. Pre-operative gabapentin has been shown to be associated with respiratory depression during anaesthesia recovery following both elective arthroplasty and laparoscopic surgical procedures.Reference Cavalcante, Sprung, Schroeder and Weingarten24,Reference Weingarten, Hawkins, Beam, Brandt, Koepp and Kellogg25 A large population-based nested case–control study found that patients receiving both prescription opioids and gabapentin had a substantial increase in the risk for opioid-related death compared with patients receiving opioids alone.Reference Gomes, Juurlink, Antoniou, Mamdani, Paterson and van den Brink9
These risks may be further exacerbated in the transoral robotic surgery population, where the oropharyngeal operation can result in airway oedema, pain requiring increased analgesia, and decreased oral intake, potentially resulting in slowed gastrointestinal motility and increased serum concentrations of analgesics. In our study, the rate of gabapentin administration was similar among those patients who did and did not have post-operative sedation. However, our protocol was conservative in regard to gabapentin dose, which may have been below the sedation threshold. Our dosing regimen could be another reason why gabapentin was not associated with decreased episodes of severe pain. The associated improvement in nasogastric tube removal and gabapentin may have been caused by drug effects (improved pain and/or anxiety with swallowing), which merits further study. However, the year of gabapentin introduction (2015) was notable for a substantial decrease in nasogastric tube duration, which raises the possibility that other practice changes may have also contributed to this reduction.
Unlike scheduled gabapentin, ibuprofen was administered as needed. The directionality of observations on pain and opioid administration is unclear from our retrospective data. Nevertheless, ibuprofen appears to be under-utilised in patients with severe pain, and scheduling ibuprofen may deliver better analgesia for these patients. Previously, ibuprofen was found to be an effective analgesic adjunct following adult tonsillectomy.Reference Attia26 However, another study found increased rates of post-operative nausea and vomiting, without opioid-sparing effects, following celecoxib administration in adult tonsillectomy patients.Reference Ng, Diamantaras, Priestley, Redman, De Silva and Mahanta27 The under-utilisation found in our study may be explained by healthcare providers choosing to use opioids at a higher rate in the early post-operative period. The historical stigma and training bias suggesting that ibuprofen may increase the risk of bleeding may also be contributing factors. Ibuprofen and other NSAIDs have been studied extensively in paediatric tonsillectomy patients, and their use has not been found to increase haemorrhage risk.Reference Bedwell, Pierce, Levy and Shah13–Reference Pfaff, Hsu and Chennupati16 Unfortunately, similar studies have not been repeated in transoral robotic surgery patients, who have more extensive surgical defects and thus experience higher rates of post-operative haemorrhage than children (5 per cent vs 3 per cent).Reference Pollei, Hinni, Moore, Hayden, Olsen and Casler12 One study did show that ibuprofen was associated with increased incidence and severity of bleeding in adult tonsillectomy patients.Reference Attia26 While we saw no association between ibuprofen use and increased haemorrhage rates, the risk cannot be determined with this limited number of patients. Ibuprofen use did not impact nasogastric tube duration, but its as-needed dosing regimen makes analysis of this outcome difficult.
Finally, it is important to understand risk factors for severe pain and sedation following transoral robotic surgery in order to improve analgesia regimens. We found an increased risk for pain with younger age, which is consistent with the literature.Reference Macintyre and Jarvis28 Male sex, typically a risk factor for increased pain,Reference Ready29 was non-significant in this cohort. This is likely because our cohort was predominantly male (90 per cent). Increasing T stage was the only variable associated with severe pain scores. In fact, increasing markers for more extensive operations, including surgical duration, tumour size, T stage, tumour subsite, intra-operative pharyngotomy and the need for bilateral neck dissections, did not result in an increase in estimated oral morphine equivalents on post-operative days 1 or 2. In addition, no patient, tumour or surgical variables were associated with sedation. The presence of a nasogastric tube at discharge was associated with an increased risk of nausea, necessitating anti-emetic management both during hospitalisation and at discharge. This should give providers pause when recommending higher doses of narcotics for patients with more ‘extensive’ operations, and suggests that titrating opioids to individual pain levels may be more effective.
• Transoral robotic surgery offers patients with human papilloma virus associated oropharyngeal squamous cell carcinoma (SCC) good long-term prognosis
• Opioid analgesics are often used to control pain following transoral robotic surgery, but can be associated with dependence and death
• Gabapentin provides modest post-operative analgesia and opioid-sparing effects, but has been associated with increased risk of death when co-prescribed with narcotics
• Non-steroidal anti-inflammatory drugs have post-operative opioid-sparing properties, and appear to be safe following tonsillectomy
• Current literature does not address optimal analgesic regimens for transoral robotic surgery patients
• This study investigated multimodal analgesia and post-operative pain in patients undergoing transoral robotic surgery for oropharyngeal SCC
Limitations
This study has several important limitations. It is retrospective in nature, and adherence to the multimodal analgesia protocol was variable. The use of gabapentin was not adopted by all surgeons, and the retrospective format of this study precludes us from ascertaining the reasons for this abstinence. Other variables in peri-operative care have changed during our study, such as the trend towards earlier nasogastric tube removal and our increasing comfort with using NSAIDs for transoral robotic surgery patients. These effects may interfere with interpreting the impact of analgesia regimens. This study focuses on post-operative days 1 and 2, without consideration of intra-operative anaesthetics, which may impact post-operative pain and sedation.
It is important to emphasise that this study cannot be used to establish safety of either gabapentin or ibuprofen in transoral robotic surgery patients, because, even though it represents a relatively large cohort, it is too small to establish risk with rare events.
Conclusion
In patients undergoing transoral robotic surgery for oropharyngeal SCC, post-operative low-dose gabapentin was not associated with measures of improved pain control or increased post-operative sedation. Possible explanations include the low to moderate dosing used in this protocol or a lack of efficacy in this patient population. Caution should be taken when considering increasing gabapentin dosing in conjunction with opioid use, and when considering pre-operative dosing, given the documented increased risk in post-operative respiratory depression. The substantial decrease in nasogastric tube duration associated with gabapentin use may either reflect efficacy of the medication in improving swallowing or be a surrogate for a clinical practice change. Further study is needed to better understand this association.
The as-needed ibuprofen administration did not result in reduction of post-operative pain. However, this medication may have been under-utilised, as it was administered less often in patients with severe pain. The likely explanation is that providers administered as-needed opioids preferentially over ibuprofen for severe pain episodes. Fortunately, increased rates of drug-specific adverse events (sedation with gabapentin, haemorrhage with ibuprofen) were not observed for either medication.
Competing interests
None declared
