Necessity of pharyngeal anesthesia during transoral gastrointestinal endoscopy: a randomized clinical trial

Article information

Clin Endosc. 2023;56(5):594-603

Publication date (electronic) : 2023 April 12

doi : https://doi.org/10.5946/ce.2022.182

Tomoyuki Hayashi^,1, Yoshiro Asahina¹, Yasuhito Takeda¹, Masaki Miyazawa¹, Hajime Takatori¹, Hidenori Kido¹, Jun Seishima¹, Noriho Iida¹, Kazuya Kitamura¹, Takeshi Terashima¹, Sakae Miyagi², Tadashi Toyama², Eishiro Mizukoshi¹, Taro Yamashita¹

¹Department of Gastroenterology, Kanazawa University, Kanazawa, Japan

²Innovative Clinical Research Center, Kanazawa University, Kanazawa, Japan

Correspondence: Tomoyuki Hayashi Department of Gastroenterology, Kanazawa University Hospital, 13-1, Takara-machi, Kanazawa, Ishikawa 920-8641, Japan E-mail: hayasix0917@gmail.com

Received 2022 July 5; Revised 2022 July 24; Accepted 2022 August 3.

Abstract

Graphic abstract

INTRODUCTION

Patients usually experience severe discomfort during transoral upper gastrointestinal endoscopy (UGE) because of the pain and strong gag reflex when the endoscope passes through the pharynx.1-3 Topical pharyngeal anesthesia, such as lidocaine (spray or viscous solution), reduces transoral UGE-induced pain and the gag reflex.4-7

According to a survey report, 0.0028% of UGE-related complications were due to pretreatment, whereas 0.00005% was fatal. The culprit drugs were sedative/analgesic-related and pharyngeal anesthesia-related in 46.5% and 8.3% of the cases, respectively.8 Thus, sedation/analgesic-related adverse events were more common than those related to pharyngeal anesthesia. However, sedation alleviates anxiety and discomfort in patients and improves their receptiveness and satisfaction with UGE.9 Although the frequency of side effects of lidocaine is lower, anaphylactoid reactions and lidocaine poisoning caused by increased blood concentration resulting in death have been reported.10 Therefore, the side effects of lidocaine are not negligible. Several studies on the use of pharyngeal anesthesia under sedation with propofol have reported that pharyngeal anesthesia does not reduce the amount of propofol, improve endoscopists’ satisfaction, or improve pharyngeal discomfort or patient satisfaction.11-14 However, at present, benzodiazepines are more commonly administered than propofol to obtain optimal sedation during UGE,9 as they do not require management by an anesthesiologist.

Therefore, the necessity of pharyngeal anesthesia in UGE under sedation with benzodiazepines is controversial. As reported previously, pharyngeal anesthesia before UGE improved endoscopy ease and patient tolerance.3 However, the number of cases was small in each study and the sedation method was not consistent. With advances in endoscopic instruments, narrow-band imaging has improved early superficial pharyngeal carcinoma detection and resulted in a better prognosis.15-17 Therefore, the importance of pharyngeal observation is increasing. However, endoscopic pharyngeal observation is difficult; careful observation of the pharynx while avoiding the gag reflex is essential to reduce the incidence of undetected cancers.

Experienced doctors can observe 86% to 90% of the sites without blind spots.18,19 However, there are no reports on whether pharyngeal anesthesia affects pharyngeal observation in patients sedated with benzodiazepines. Therefore, we conducted a prospective, single-blinded, randomized clinical trial to compare the groups undergoing transoral endoscopy with and without pharyngeal anesthesia under sedation with benzodiazepines. In addition to evaluating the degree of pain, we assessed pharyngeal observation quality between the two groups. This study aimed to prove the hypothesis that pharyngeal anesthesia is not required in patients sedated with benzodiazepines. If the group not receiving pharyngeal anesthesia was not inferior to the group receiving it, pharyngeal anesthesia was considered unnecessary. Therefore, a non-inferiority trial was deemed appropriate for testing this hypothesis. Randomized controlled trials evaluating patients’ pain and pharyngeal observation ability with pharyngeal anesthesia compared to those without pharyngeal anesthesia under sedation with benzodiazepines are required.

METHODS

Study design

This single-center, prospective, single-blinded, randomized study was conducted at Kanazawa University Hospital, Ishikawa, Japan. Patients were randomized into groups, with or without pharyngeal anesthesia, using computer-generated numerical codes by the information manager (simple randomization, allocation ratio 1:1).

After adopting the pharyngeal anesthesia protocols described below, the first image upon insertion of the endoscope was captured, followed by six and four endoscopic images of the oropharynx and hypopharynx, respectively. The last image was captured before insertion into the esophagus. Overall, 12 narrow-band imaging images were acquired for each patient. We also measured the time from the first image to insertion into the esophagus. Once it was impossible to obtain images owing to the continuous gag reflex, the endoscope was inserted into the esophagus. Hence, the first and last images were used only for time measurement.

The physicians were required to undergo more than half a year of pharyngeal observation training prior to participating in this study. Two endoscopists (TH and YA) assessed whether the 10 images were appropriate (Fig. 1). They were blinded to the allocation group and independent of the examining physicians. According to previous reports,18,19 the definition of an appropriate image requires the fulfillment of three criteria. First, an image was captured at an appropriate location. Second, we focused on the images. Third, the mucus was removed, and the pharyngeal mucosa color in the images was evaluated. Consensus was reached upon discussion among expert endoscopists in cases of disagreement.

Fig. 1.

The 10 images of the oropharynx and hypopharynx. The definition of a high-quality image required three criteria: (1) the image was taken at the appropriate location; (2) the image was focused; and (3) mucus had been removed, and it was possible to evaluate the color of the pharyngeal mucosa.

The following data were recorded for each patient after endoscopy: age, sex, body mass index (BMI), performance status (PS) of the Eastern Cooperative Oncology Group, drinking and smoking habits, and the number of endoscopic experiences. Subjective pain symptoms were evaluated using an 11-point visual analog scale (VAS), with 10 points indicating the most severe pain and zero indicating no pain. Data were collected using a questionnaire after they were completely awake.

RESULTS

Figure 2 shows a flow chart of patient selection and allocation. A total of 500 patients were enrolled and randomly allocated to the PA– and PA+ groups; 250 in each group. The patient demographics and clinical characteristics are shown in Figure 3. The number of patients in both groups was similar. No statistically significant differences were observed in age, sex, height, weight, BMI, PS, drinking and smoking habits, number of endoscopic procedures, and the Ramsay score at sedation.

Fig. 2.

Flow chart of patient enrollment. UGE, upper gastrointestinal endoscopy; PA, pharyngeal anesthesia.

Fig. 3.

Patients demographics and clinical characteristics and subgroup analysis of patients’ characteristics. CI, confidence interval.

The agreement rate between the two endoscopists was 97.2% and the kappa statistic was 0.916.

The pharyngeal observation success rate in the PA– group was 72.0% (95% confidence interval [CI], 66.4%–77.6%), whereas that in the PA+ group was 84.0% (95% CI, 79.5%–88.5%). The difference was –19.2% to –4.8% (p=0.707, non-inferiority), and the non-inferiority of the PA group was not demonstrated (Table 1). In the PA– group, one oropharyngeal cancer, and one hypopharyngeal cancer were detected. Table 2 shows the differences in situations during endoscopy. The PA– group was inferior to the PA+ group in the number of observable parts (0–10) (8.33±2.32vs. 8.86±1.94, p=0.006), pharyngeal observation time (67.2±33.7 vs. 58.2±27.1 seconds, p=0.001), pain evaluated by VAS (1.21±2.37 vs. 0.68±1.78, p=0.004), and in the number of gag reflexes (3.83±3.15 vs. 2.11±2.49, p<0.001). There were no significant differences in adverse events, decrease in SpO₂ (12.0% vs. 10.0%, p=0.475), decrease in blood pressure (1.2% vs. 0.4%, p=0.623), or bradycardia (0.4% vs. 1.2%, p=0.623) between the two groups.

Table 1.

Non-inferiority test of the difference of pharyngeal observation success rate (based on a risk α of 0.025 [one-sided test], a statistical power of 80%, and a non-inferiority threshold [Δ] of 0.1)

Table 2.

Differences in pharyngeal observation ability, pain, and adverse events during endoscopy

Table 3 shows the percentages of suitable quality images obtained at the 10 prescribed points. The percentage of suitable quality images at the posterior wall of the oropharynx (92.4% vs. 96.8%, p=0.030), vocal fold (74.8% vs. 88.8%, p<0.001), and right (74.4% vs. 88.4%, p<0.001) and left pyriform sinuses (76.4% vs. 88.0%, p=0.001) were significantly lower in the PA– than in the PA+ group. No significant differences were observed at any of the other parameters.

Table 3.

Percentages of suitable quality images obtained at the 10 tested points

Subgroup analysis of the lower level of sedation (Ramsay score <5 in 217 patients) revealed that the pharyngeal observation success rate (61.6% vs. 81.9%, p=0.001), number of pharyngeal observation sites (7.88±2.55 vs. 8.89±1.89, p=0.001), and number of gag reflexes (3.97±3.07 vs. 2.70±2.73, p=0.001) were significantly lower in the PA– than in the PA+ group. However, non-significant differences were found in pain evaluated by VAS (1.85±2.91 vs. 1.27±2.42, p=0.113) and pharyngeal observation time (67.0±32.9 vs. 59.8±31.9 seconds, p=0.104) (Table 1). The subgroup analysis of the higher level of sedation (Ramsay score ≥5 in 283 cases) revealed that pain evaluated by VAS (0.70±1.66 vs. 0.25±0.77, p=0.005), pharyngeal observation time (67.3±34.4 vs. 56.7±23.1 seconds, p=0.003), and the number of gag reflexes (3.70±3.22 vs. 1.68±2.20, p<0.001) were significantly lower in the PA– than in the PA+ group. However, a non-significant difference between the two groups was found in the pharyngeal observation success rate (80.4% vs. 85.5%, p=0.255) and the number of pharyngeal observation sites (8.70±2.05 vs. 8.83±1.97, p=0.562) (Table 1). Figure 3 shows the subgroup analysis of patients’ characteristics, such as sex, age (over or under 70), BMI (over or under 25), alcohol intake (over or under one time/week), smoking status, PS (0 or >1), antihypertensive or psychotropic drugs use, Ramsay score (over or under 5), and the number of examinations (over or under 5). In comparing the BMI and Ramsay scores, the p values were <0.05.

DISCUSSION

To the best of our knowledge, this is the first study to evaluate pharyngeal observation ability in patients undergoing transoral UGE under sedation with or without administration of topical pharyngeal anesthesia. Our study revealed that non-inferiority of non-pharyngeal anesthesia was not demonstrated in terms of pharyngeal observation success rate, and the percentage of suitable quality images at the posterior wall of the oropharynx, vocal fold, and pyriform sinus was significantly inferior. However, in the subgroup analysis of higher levels of sedation, there were almost no differences in the pharyngeal observation success rate between patients with and without pharyngeal anesthesia.

However, in the ad hoc analysis, the PA+ group had a significantly higher pharyngeal observation success rate than the PA– group (p=0.001). Naturally, it is not a preset test; however, the PA+ group may have better pharyngeal observation ability than the PA– group for patients who were administered intravenous anesthesia with midazolam.

The PA+ group was significantly superior to the PA– group in terms of the number of observable pharyngeal parts (0.53 parts), observation time (9.0 seconds), pain measured using VAS (0.53 points), and the number of gag reflexes (1.72 times) owing to the advantages of pharyngeal anesthesia using lidocaine. These differences were presumably small absolute numerical values between the groups. However, these are clinically important facts for endoscopists, who focus on reducing patients’ pain as much as possible. Pharyngeal anesthesia may be tolerated by weighing these differences against side effects.

In the subgroup analysis of each pharyngeal site, the PA– group was statistically inferior to the PA+ group in terms of observation of the posterior wall of the oropharynx, epiglottis, vocal fold, and right and left pyriform sinuses. This difference can be attributed to multiple comparisons. It is suspected that the effect of anesthesia on the hypopharynx may be stronger in the PA+ group; however, more cases need to be evaluated to prove the theory. Further studies on the relationship between the extent of hypopharyngeal anesthesia and pharyngeal observation ability are desirable because the pyriform sinus is the predilection site for pharyngeal cancer.

In the subgroup analysis for each variable, obese patients (BMI ≥25 kg/m²) and patients with low sedation levels (Ramsay score <5) may have had low pharyngeal observation success rates in the PA group. To clarify these differences, it is necessary to include a larger number of cases.

Detailed subgroup analysis of each sedation level revealed that the pharyngeal observation success rate and the number of observable sites were low in the PA group at low sedation levels (Ramsay score <5). However, these differences were not observed at high sedation levels (Ramsay score ≥5). Based on our findings, it is reasonable to presume that deep sedation reduces the disadvantage of not using pharyngeal anesthesia. This would be beneficial for patients who cannot use pharyngeal anesthesia because of allergies. Although the possibility of chance cannot be ruled out by multiple analyses, as deep sedation with propofol has shown that pharyngeal anesthesia is unnecessary, the deeper the sedation level, the lesser the pharyngeal anesthesia that might be required. No data have clarified this hypothesis. Further studies based on this hypothesis are required.

In the deep sedation group, pharyngeal anesthesia may further reduce pain and pharyngeal observation time. Although difficult to explain theoretically, this may be because pharyngeal anesthesia reduces pain and the difficulty in observation. Alternatively, considering that there is a similar tendency without significant difference in the low sedation group, it may not be relevant to sedation. Large-scale studies are necessary to further investigate this.

Diagnostic transnasal endoscopy has been proven to be more tolerable than the transoral approach.24,25 Furthermore, during the coronavirus disease 2019 pandemic, transnasal endoscopy has proven to be beneficial in terms of less aerosol spreading.26 However, transoral endoscopy models have magnification and better image quality. Therefore, endoscopists prefer this approach.

This study had some limitations. First, it was conducted at a single institution. Second, it was single-blind because the patients knew whether pharyngeal anesthesia was administered. The PA group did not receive a placebo intervention because of the difficulty of creating a placebo with a flavor similar to that of lidocaine. Thus, the participants’ awareness of whether pharyngeal anesthesia was administered might have altered their mental state and affected the outcome. Third, our study did not evaluate the postcricoid area. The nasal Valsalva or sniffing position has been reported to improve examination in transoral endoscopy27,28; however, we did not adopt such methods. However, further studies are required to explore these methods. Fourth, no synthetic opioid analgesics such as pethidine were used as premedication. However, our findings are important for facilities that are unable to use them. Fifth, high-risk patients with pharyngeal cancer often have a history of alcohol consumption and may tolerate sedation. Such patients are considered more suitable for participation. Sixth, the definition of successful pharyngeal observation formulated by us may seem controversial among experts. The results of this study were based on two expert endoscopists’ opinions of endoscopic images, and the outcomes were based on subjective evidence without validation. To date, there has been no consensus among experts in this regard. Therefore, it is desirable to establish clear criteria for future clinical studies.

In conclusion, we found that not administering pharyngeal anesthesia did not demonstrate non-inferiority of pharyngeal observation ability. Pharyngeal anesthesia may improve pharyngeal observation ability, especially in the hypopharynx, and reduce patients' pain. However, this difference may be reduced by deeper anesthesia.

Article information Continued