Determination of the ED50 of Tegileridine and Comparison with Fentanyl for Analgesia During Hysteroscopy in Chinese Female Patients: An Exploratory Sequential Dose‑Finding Study Authors: Houren Xu, Feng Lu, Yu Wang, Xinxin Ruan, Haiyu Xie, Maolin Zhong Published: 2026 ABSTRACT ObjectiveTegileridine is a novel selective μ-opioid receptor agonist. This exploratory dose‑finding study aimed to estimate the 50% effective dose (ED50) and secondarily the 95% effective dose (ED95) of tegileridine for suppressing the cervical dilation reflex during hysteroscopic surgery under propofol anesthesia, compare it with fentanyl, and evaluate perioperative safety in an exploratory manner.MethodsThis prospective, randomized, single‑center trial enrolled 54 patients (ASA I‑II). Patients received tegileridine (initial dose 16.67 μg/kg) or fentanyl (initial dose 2 μg/kg) with propofol. Doses were adjusted using the Dixon sequential method. The primary endpoint was ED50 with its 95% confidence interval (CI), estimated using the percentile bootstrap method (2,000 replicates) as the primary inferential framework. ED95 was calculated secondarily via probit regression. Safety data are reported descriptively.ResultsTwenty‑seven patients per group completed the study. The bootstrap‑based ED50 for tegileridine was 23.22 μg/kg (95% CI: 22.39‑24.19); for fentanyl, 1.51 μg/kg (95% CI: 0.97‑4.07). The ED50 potency ratio (tegileridine:fentanyl) was approximately 1:15 (95% uncertainty range derived from bootstrap CIs: 1:5.5 to 1:24.9). The ED95 values from probit regression were 23.89 μg/kg (95% CI: 23.68‑24.12) for tegileridine and 1.97 μg/kg (95% CI: 1.87‑2.07) for fentanyl. Respiratory depression/hypoxemia was observed in 14.8% of the tegileridine group and 59.3% of the fentanyl group; these exploratory findings should be interpreted with caution given the small sample and open‑label design.ConclusionThis exploratory study estimated the ED50 of tegileridine and fentanyl for hysteroscopic surgery. The ED50-based potency ratio and the observed differences in safety outcomes require confirmation in larger, adequately powered, blinded trials. The ED95 estimates are model‑based and carry substantial uncertainty.Trial RegistrationChinese Clinical Trial Registry, ChiCTR2500107995. Introduction Hysteroscopic surgery is currently considered to be one of the gold standards for diagnosing and treating intrauterine pathologies.1–3 During the course of outpatient hysteroscopic anesthesia management, acute pain stimuli resulting from cervical dilation and traction are particularly salient. These often manifest as body movements, breath-holding, or reflexive circulatory fluctuations, thereby increasing the difficulty of endoscopic manipulation and compromising the smooth execution of the procedure.4,5 Consequently, the effective suppression of pain perception and body movements during cervical dilation, whilst ensuring rapid recovery, remains a critical challenge in clinical anesthesia practice. At present, the combination of propofol with opioid agents via intravenous anesthesia is a widespread practice for outpatient hysteroscopic procedures.6 However, traditional μ-opioid receptor agonists such as fentanyl, while mediating analgesia through G protein pathways, also significantly activate the β-arrestin-2 pathway. This has been shown to result in dose-limiting adverse effects, such as respiratory depression and nausea/vomiting, which collectively constitute a clinically significant “opioid burden”.7,8 It is important to note that even in short-duration procedures such as hysteroscopy, the administration of a single dose of fentanyl carries a risk of perioperative respiratory depression that should not be overlooked. Consequently, current international guidelines and consensus statements advocate the exploration of safer analgesic strategies, such as the utilisation of μ-opioid receptor partial agonists, with a view to minimizing opioid-related adverse events, particularly respiratory depression.9,10 Tegileridine fumarate is a novel μ-opioid receptor partial agonist pharmacologically engineered to selectively activate the G protein pathway for analgesic effects while minimizing activation of the β-arrestin-2 pathway, theoretically reducing the risk of associated adverse reactions.7,11 A body of research has emerged from preclinical studies and early clinical trials that suggest the medication is effective in providing analgesia and has the potential to be safe.12–14 The results of clinical research on the comparable drug oliceridine suggest that the biased mechanism may reduce the risk of respiratory depression while providing equivalent analgesia.15–17 Nevertheless, the pharmacodynamic and safety profiles of different biased agonists still require separate validation in specific surgical settings. Despite the expansion of clinical research on tegileridine in terms of patient populations and surgical types, the availability of dose-response data using cervical dilation-induced movement as a key clinical endpoint remains limited in short-duration procedures such as hysteroscopy. The accurate determination of population-specific effective doses, particularly the 95% effective dose (ED95) which covers the majority of patients, is fundamental for guiding clinical dosing decisions and achieving precision anesthesia and analgesia. The objective of this exploratory study was to estimate the ED50 and, secondarily, the ED95 of tegileridine fumarate for suppressing cervical dilation‑induced motor response under propofol‑based anesthesia, and to compare its dose‑response relationship with that of fentanyl. Given the well‑known limitation that the Dixon up‑and‑down method is optimized for ED50 rather than ED95, we prespecified ED50 as the primary statistical endpoint. Safety and efficacy data are reported as exploratory findings. Materials and Methods Study Design and Ethics The present study constitutes a prospective, randomized, single-center, dose-exploratory trial employing a modified Dixon sequential (up-and-down) method to investigate the dose-response relationship of tegileridine fumarate versus fentanyl in inhibiting cervical dilation-induced motor responses across two independent sequences. This study complies with the Declaration of Helsinki. The study protocol was approved by the Ethics Committee of The First Affiliated Hospital of Gannan Medical University (Approval No.: 22sc-2025 No. 178) and registered with the China Clinical Trial Registry (Registration No.: ChiCTR2500107995). Prior to enrolment, written informed consent was obtained from all participants. Study Population The study enrolled patients scheduled for hysteroscopic day surgery at our hospital between August 2025 and December 2025. The inclusion criteria were as follows: (1) Age 18–65 years; (2) American Society of Anesthesiologists (ASA) physical status classification of I–II; (3) Body Mass Index (BMI) 18–28 kg/m2. The exclusion criteria were as follows: (1) Known allergy to the study drug or any of its components; (2) Severe cardiac, hepatic, or renal insufficiency; (3) Difficult airway assessment or recent respiratory tract infection; (4) Pregnant or lactating women; (5) Chronic opioid use or history of substance abuse; (6) Risk of gastrointestinal obstruction. During the study period, 58 patients scheduled for day-case hysteroscopic surgery were screened. Four patients were excluded from the study due to non-compliance with the inclusion criteria (n = 3) or due to their refusal to participate (n = 1). In conclusion, 54 patients were successfully enrolled in the study and all completed the trial process, thus forming the per-protocol set for analysis. Anesthesia Protocol and Sequential Design Randomization Patients were randomly assigned in a 1:1 ratio to the tegileridine group (T) or the fentanyl group (F) using a computer-generated random number sequence (R software 4.3.2). Due to the different administration protocols required for the two drugs (tegileridine via 10-minute intravenous infusion pump, fentanyl via 90-second intravenous bolus injection), the anesthesiologists administering the drugs could not be blinded to group allocation. However, to minimize observation bias, the researchers responsible for the primary endpoint assessment, data entry, and statistical analysis remained blinded to group assignments. The primary endpoint assessment was conducted by evaluators who were unaware of the treatment groups during the cervical dilation phase and prior to the administration of any rescue analgesia. Anesthesia Administration Preoperative Preparation and Monitoring All patients undergo a standard preoperative fasting period of 8 hours and a 2-hour restriction on fluid intake. Upon entering the operating theatre, preoxygenation is administered via face mask (FiO2 100%, 3 L/min). The standard monitoring protocol encompasses a range of non-invasive assessments, including blood pressure, electrocardiogram, pulse oximetry (SpO2), end-tidal carbon dioxide (EtCO2), measured via nasal cannula microflow sampling, and bispectral index (BIS). The target intraoperative BIS range is maintained between 40 and 60. Anesthetic Induction and Initial Drug Administration Following intravenous bolus injection of propofol (2 mg/kg) to complete anesthetic induction, study drugs were administered according to group assignment: Group T (tegileridine Group): The intravenous infusion of an initial dose of 16.67 μg/kg tegileridine fumarate (diluted with 0.9% sodium chloride to 20 mL and infused at a constant rate over 10 minutes).13 The F Group (Fentanyl Group): The initial dose of 2.0 μg/kg fentanyl was administered intravenously as a bolus (diluted in 0.9% sodium chloride to 10 mL and slowly injected over 90 seconds). Timing of Surgical Initiation In order to ensure optimal consistency within the therapeutic exposure window, it was imperative that cervical dilation procedures were initiated only subsequent to the fulfillment of the following criteria: (1) disappearance of the ciliary reflex; (2) MOAA/S score of 0–1;18 (3) Based on the aforementioned conditions, the timing of surgical initiation was determined on a case-by-case basis for each group according to their respective pharmacokinetic characteristics: Tegileridine group: The initiation of the procedure occurred immediately after the completion of the intravenous infusion, which had a duration of 10 minutes. The design of the study is consistent with the pharmacokinetic profile of the substance, which is such that peak plasma concentrations are reached upon completion of the infusion, thereby enabling immediate analgesic onset.19 Fentanyl group: The initiation of the process occurred a mere four minutes after the completion of the intravenous bolus injection, with a duration of 90 seconds. The duration of the waiting period was determined by pharmacodynamic studies indicating that the maximum respiratory depression effect of fentanyl (a surrogate for potent analgesia) peaks approximately five minutes (median) post-administration following a single intravenous bolus.20 Subsequent to the fulfillment of the aforementioned conditions, a gynecologist, uninformed of the group assignments, undertook standardized mechanical cervical dilation employing Hegar dilators, commencing with size 4 and successively increasing sizes until size 8 dilator traversed without impediment. The present study defined dilation to size 7.5 as the primary pain stimulus observation point. Anesthesia Maintenance and Rescue Analgesia The sedation of the patient was maintained throughout the surgical procedure by the continuous intravenous infusion of propofol (4–8 mg/kg/h). In instances where the MOAA/S score exceeded 1 point, an additional 0.5 mg/kg of propofol was administered. The primary efficacy endpoint of this study was defined as the occurrence of motor response during initial cervical dilation (to 7.5 cm). Motor activity was assessed by an independent investigator according to predefined criteria: The term “failure” (ie. inadequate analgesia) was defined as the occurrence of distinct trunk or limb movements that impeded dilation or necessitated interruption. Conversely, the term “success” was defined as the absence of such movements. This evaluation of success/failure was conducted prior to the administration of any rescue medication. In instances where failure was identified, a single rescue dose of the study drug was administered intravenously (T group: tegileridine 1.67 μg/kg; F group: fentanyl 0.5 μg/kg) to ensure safe and uninterrupted surgical progression. Sequential Design Sequential Dose Adjustment Rule (Modified Dixon Up-and-Down Method) The present study employed a modified Dixon sequential design for dose escalation, with the following specific rules: Response Assessment: In the context of a single patient, the absence of body movement during cervical dilation is designated a “negative response (success)”. Conversely, the presence of body movement is categorized as a “positive response (failure)”. Dose Adjustment: The subsequent patient’s dosage is contingent upon the response exhibited by the preceding patient within the same cohort. In the event that the preceding patient exhibited a “positive response (failure)”, the dosage is to be augmented by a preset increment. In the event that the preceding patient exhibited a “negative response (success)”, the dosage is to be reduced by one preset step. Step Size: The step size for dose adjustment is set as follows: The dosage of tegileridine was set at 1.67 μg/kg (approximately 10% of the initial dose), and that of Fentanyl at 0.2 μg/kg (approximately 10% of the initial dose). The rationale underpinning this step size design is to ensure clinically meaningful dose adjustments while avoiding excessive exposure to ineffective or excessive doses due to overly large increments, or prolonged sequential processes due to overly small increments.21 Sequential Termination Criteria The sequential test is continued until each group achieves a minimum of seven consecutive “failure-success” reversals (ie., crossover points). At this juncture, the dose distribution is deemed stable and the sequential phase is brought to a conclusion. This termination criterion aligns with the recommendations of the classical Dixon methodology, ensuring a stable estimate of the ED50 for the study population.21 Data Collection Baseline data: Collect baseline information for all enrolled patients, including their demographic details, their ASA physical status classification and their BMI. Timeline data: Record key time points, including the start of anesthesia induction (T0), the start of surgery (defined as the initiation of cervical dilation, T1) and the end of surgery (T2). Vital signs: Continuously monitor and record heart rate (HR), non-invasive blood pressure (systolic SBP, diastolic DBP and mean arterial pressure MAP) and pulse oximetry (SpO2) at T0, T1 and T2. Primary Endpoint: Document the physical response during cervical dilation (successful/unsuccessful). Drug dosage: Record the actual doses administered of the study drugs (tegileridine fumarate and fentanyl) and total propofol consumption. Safety monitoring: Record any adverse events that occur during surgery, paying particular attention to respiratory depression (defined as an oxygen saturation level of less than 92% or a respiratory rate of less than eight breaths per minute) and hypotension (defined as a decrease in mean arterial pressure of more than 20% from baseline). Study Endpoints Primary endpoint: ED50 values and their 95% confidence intervals for cervical dilation inhibition in both groups estimated using percentile bootstrap resampling as the primary inferential framework. Key secondary endpoints: ED95 values (model‑based, via probit regression); dose‑response curve shape parameters; exploratory safety outcome. Exploratory endpoint: incidence of perioperative adverse events. Sample Size Calculation The sample size for this study was determined based on the design principles of the modified Dixon sequential method. According to the classical literature on this method, enrolling 20 to 30 subjects per group is typically sufficient when the dose adjustment steps are appropriately set to obtain a stable estimate of the ED50.21 Furthermore, to ensure stable dose distribution and meet the fundamental requirements for sequential analysis, the study prespecified a termination criterion of obtaining at least seven consecutive “failure-to-success” reversals per group.21 Based on these principles, the sample size was set at 27 subjects per group, totaling 54 subjects. Statistical Analysis This study strictly adhered to the design and analysis specifications of the modified Dixon sequential (up-and-down) method, with the relevant methodological details referenced in the authoritative review by Pace and Stylianou.21 The dose escalation process, response definitions and termination criteria were all specified in advance and rigorously enforced throughout the trial to ensure the objectivity and reproducibility of the dose exploration. Statistical Analysis Software All statistical analyses were performed in the R 4.3.2 environment. The primary packages used were drc (version 3.0–1) for dose-response curve fitting, boot (version 1.3–30) for bootstrap confidence interval estimation, ggplot2 (version 3.5.1) for visualising the data, and dplyr for manipulating the data. Descriptive Statistical Analysis Baseline characteristics are reported descriptively using medians (interquartile range, IQR) for continuous variables and frequencies (percentages) for categorical variables. No hypothesis tests were performed for baseline comparisons. For outcome comparisons, Fisher’s exact test was used for categorical variables, and the Wilcoxon rank-sum test for continuous variables, as appropriate. All tests were two-sided, and a P-value <0.05 was considered statistically significant for the primary and secondary outcomes. Given the exploratory nature of this dose-finding study, no adjustment for multiple comparisons was applied. Dose-Response Analysis The ED95 values and their 95% confidence intervals for the inhibition of cervical dilation by both drugs were calculated using a probit regression model. To characterize the dose-response relationship comprehensively, the dose-response curve was fitted using a four-parameter logistic (4PL) nonlinear regression model. ED50 Estimates Were Obtained from the Four-Parameter Logistic Model as Described in Confidence Interval Estimation Methods Safety Data Analysis Fisher’s exact test was used to compare the incidence rates of adverse events between groups. To quantify the difference in safety between tegileridine and fentanyl further, the relative risk (RR), risk difference (RD) and their respective 95% confidence intervals were calculated. An RR less than 1 indicates a lower risk in the tegileridine group. Additionally, the number needed to treat (NNT) was calculated, representing the number of patients who would need to be treated with tegileridine to prevent one adverse event compared to fentanyl. Confidence Interval Estimation Methods The primary analysis for ED50 estimation used percentile bootstrap resampling with 2,000 replicates, as this approach provides a unified inferential framework without distributional assumptions. Point estimates of ED50 were obtained from a four‑parameter logistic (4PL) model. The 95% confidence intervals for ED50 were derived from the bootstrap percentiles. Probit regression was used as a complementary sensitivity analysis for ED50 and as the main method for ED95 (secondary). For the fentanyl group, a truncated bootstrap (retaining only positive estimates) was applied to avoid negative lower bounds due to the small sample size; this does not affect the primary bootstrap‑based inference. The ED95 and its 95% CI were calculated using probit regression and are presented as model‑based exploratory estimates. Visual Analysis All graphics were generated using the ggplot2 package in R, following scientific visualization principles. The figures include dose-response scatter plots with fitted curves, forest plots of parameter estimates, dose distribution histograms, and response rate bar charts. Sensitivity Analysis A sensitivity analysis was conducted on the bootstrap confidence interval estimates for ED50 to assess the robustness of the analytical results. This analysis compared the differences in the results obtained using the standard percentile method and the truncation method (the latter of which was applied only to the fentanyl group). Results Patient Flow and Baseline Characteristics A total of 58 patients were screened, of whom 54 (27 in each group: tegileridine® and fentanyl) completed the trial and were included in the analysis. The processes of patient screening, randomisation, allocation and analysis are detailed in Figure 1 (CONSORT flow diagram). The two groups were comparable with respect to baseline characteristics (Table 1). No statistical comparisons were performed.Table 1Baseline Characteristics of Patients in the Dose-Finding PhaseVariableTegileridine (n = 27)Fentanyl (n = 27)Demographic dataAge, y47.0 (38.0–53.5)43.0 (38.0–49.0)BMI, kg/m221.2 (20.5–24.4)22.6 (21.1–24.8)ASA classification, n (%)ASA I20 (74.1)19 (70.4)ASA II7 (25.9)8 (29.6)Obstetric historyVaginal delivery, n (%)23 (85)22 (81)Surgical dataSurgery duration, min13.8 (11.6–19.2)13.2 (11.8–18.6)Propofol consumption, mg228.0 (200.0–261.2)240.5 (207.8–280.0)Note: Data are presented as median (interquartile range) or n (%).Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index. Figure 1CONSORT flow diagram. Patient enrollment, randomization, allocation, and analysis in the dose-finding study comparing tegileridine and fentanyl for hysteroscopy.The flowchart begins with 'Assessed for eligibility (N = 58)', leading to 'Excluded (n = 4)' with subpoints 'Not meeting criteria: 3' and 'Declined: 1'. The next step is 'Randomized (N = 54)', which splits into two paths: 'Allocated to Tegileridine (n = 27)' and 'Allocated to Fentanyl (n = 27)'. Both paths continue with 'Received intervention (n = 27)', followed by 'Lost to follow-up (n = 0)' and conclude with 'Analyzed for ED95 (n = 27)'.Study flow: enrollment, randomization, allocation, analysis of tegileridine vs. fentanyl. Table 1Baseline Characteristics of Patients in the Dose-Finding PhaseVariableTegileridine (n = 27)Fentanyl (n = 27)Demographic dataAge, y47.0 (38.0–53.5)43.0 (38.0–49.0)BMI, kg/m221.2 (20.5–24.4)22.6 (21.1–24.8)ASA classification, n (%)ASA I20 (74.1)19 (70.4)ASA II7 (25.9)8 (29.6)Obstetric historyVaginal delivery, n (%)23 (85)22 (81)Surgical dataSurgery duration, min13.8 (11.6–19.2)13.2 (11.8–18.6)Propofol consumption, mg228.0 (200.0–261.2)240.5 (207.8–280.0)Note: Data are presented as median (interquartile range) or n (%).Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index. Baseline Characteristics of Patients in the Dose-Finding Phase Note: Data are presented as median (interquartile range) or n (%). Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index. CONSORT flow diagram. Patient enrollment, randomization, allocation, and analysis in the dose-finding study comparing tegileridine and fentanyl for hysteroscopy. Sequential Dose Exploration Process This study used a modified Dixon sequential method to explore and adjust the dosage. The initial doses of tegileridine and fentanyl were 16.67 μg/kg and 2.0 μg/kg, respectively, with adjustment increments of 1.67 μg/kg and 0.2 μg/kg, respectively. Dose adjustments continued until each group achieved at least seven consecutive pairs of “failure-success” reversal events (eight pairs were ultimately observed in each group). At this point, the dose distribution was considered stable and the sequential phase was concluded. Figure 2A and B show the dose adjustment sequences for tegileridine and fentanyl, along with patient responses, respectively. Figure 2Dose-response sequences determined by Dixon’s up-and-down method for (A) tegileridine and (B) fentanyl. Positive responses (no movement, success) are shown as solid triangles (▲); negative responses (movement, failure) as hollow circles (○). Connecting lines illustrate the dose adjustment trajectory based on the response of the previous patient. Initial dose: tegileridine 16.67 μg/kg (step size 1.67 μg/kg); fentanyl 2.0 μg/kg (step size 0.2 μg/kg).Image A depicts a dose sequence for Fentanyl with outcomes: negatives as open circles, positives as filled triangles. The x-axis shows consecutive patients (1-27) and the y-axis shows doses (1.6-2.4 µg/kg). Starting at 2.0, the dose drops to 1.6 by patient 3, peaks above 2.4 at patient 10, declines to 1.8 by patient 14, oscillates between 1.8-2.0 until patient 21, then trends down to 1.4 by patient 24, rising to 1.6 by patient 27. Image B illustrates a dose sequence for Tegileridine with similar markers. The x-axis is the same, while the y-axis ranges from 17.5-25.0 µg/kg. Starting near 17.0 at patient 1, the dose rises to 25.0 by patient 6, then fluctuates between 21.5-26.5, peaking at 26.5 around patient 14, with returns to 25.0 at patients 11, 15, 18 and 20.Two line graphs showing Dixon up-and-down dose sequences for fentanyl and tegileridine across patients. Dose-response sequences determined by Dixon’s up-and-down method for (A) tegileridine and (B) fentanyl. Positive responses (no movement, success) are shown as solid triangles (▲); negative responses (movement, failure) as hollow circles (○). Connecting lines illustrate the dose adjustment trajectory based on the response of the previous patient. Initial dose: tegileridine 16.67 μg/kg (step size 1.67 μg/kg); fentanyl 2.0 μg/kg (step size 0.2 μg/kg). Detailed dose‑level summaries, including number of patients treated, successes, failures, and reversal points, are presented in Table 2 (fentanyl) and Table 3 (tegileridine). The tables confirm that each group reached the prespecified termination criterion (≥7 consecutive failure‑success reversal pairs). Individual patient sequences are illustrated in Figure 2A and B.Table 2Dose-Level Summary for Fentanyl GroupDose (μg/kg)Patients TreatedSuccessesFailures1.61011.8131122.013121Total271314Note: Initial dose 2.0 μg/kg, step size 0.2 μg/kg. For reversal points, see Figure 2A. Table 3Dose-Level Summary for Tegileridine GroupDose (μg/kg)Patients TreatedSuccessesFailures16.6730318.3462420.0162421.6864223.35660Total271413Note: Initial dose 16.67 μg/kg, step size 1.67 μg/kg. For reversal points, see Figure 2B. Table 2Dose-Level Summary for Fentanyl GroupDose (μg/kg)Patients TreatedSuccessesFailures1.61011.8131122.013121Total271314Note: Initial dose 2.0 μg/kg, step size 0.2 μg/kg. For reversal points, see Figure 2A. Dose-Level Summary for Fentanyl Group Note: Initial dose 2.0 μg/kg, step size 0.2 μg/kg. For reversal points, see Figure 2A. Table 3Dose-Level Summary for Tegileridine GroupDose (μg/kg)Patients TreatedSuccessesFailures16.6730318.3462420.0162421.6864223.35660Total271413Note: Initial dose 16.67 μg/kg, step size 1.67 μg/kg. For reversal points, see Figure 2B. Dose-Level Summary for Tegileridine Group Note: Initial dose 16.67 μg/kg, step size 1.67 μg/kg. For reversal points, see Figure 2B. Following the sequential trial, Figure 3 depicts the final dose distributions of the study drugs administered to patients in both groups. The tegileridine group exhibited a typical stepwise dose distribution, whereas the fentanyl group demonstrated a more concentrated distribution. Figure 3Distribution of final doses administered to patients in both groups. (A) Fentanyl group exhibited a more concentrated dose distribution. (B) Tegileridine group demonstrated a typical stepwise dose distribution.The image A showing a histogram titled Fentanyl. The x-axis label is Dose(ug/kg). The y-axis label is Frequency. The y-axis ranges from 0 to 8 with labeled ticks at 0, 2, 4, 6 and 8. Bars are shown at approximately 1.4 with frequency 2, 1.6 with frequency 4, 1.8 with frequency 8, 2.0 with frequency 8, 2.2 with frequency 2, 2.4 with frequency 2 and 2.6 with frequency 1. The image B showing a histogram titled Tegileridine. The x-axis label is Dose(ug/kg). The y-axis label is Frequency. The y-axis ranges from 0 to 12 with labeled ticks at 0, 3, 6, 9 and 12. Bars are shown at approximately 16.5 with frequency 1, 18.0 with frequency 1, 20.0 with frequency 1, 21.5 with frequency 6, 23.5 with frequency 11, 25.0 with frequency 6 and 26.5 with frequency 1.A pair of histograms showing final dose distributions for fentanyl and tegileridine groups. Distribution of final doses administered to patients in both groups. (A) Fentanyl group exhibited a more concentrated dose distribution. (B) Tegileridine group demonstrated a typical stepwise dose distribution. Estimated Effective Doses and Potency Comparison Using sequential trial data, we constructed dose‑response curves (Figure 4). The primary analysis for ED50 estimation used percentile bootstrap resampling (2,000 replicates) as the unified inferential framework. Probit regression and four‑parameter logistic (4PL) models were employed as complementary sensitivity analyses. Figure 4Dose-response curves of tegileridine and fentanyl for inhibiting cervical dilation-induced motor responses. (A) Fentanyl dose-response curve. (B) Tegileridine dose-response curve. (C) Comparative dose-response curves of both drugs.Image A: Fentanyl Dose-Response Curve shows dose (ug/kg) vs. response probability (0.0-1.0). ED95 is 1.97 ug/kg (95% CI: 1.87-2.07); ED50 is 1.51 ug/kg (95% CI: 0.97-4.07). Horizontal lines at 0.5 and near 0.95; vertical lines at 1.51 and 1.97. Curve rises from (1.4, 0.2) to (2.6, 0.57). Points at y=0 and y=1 across doses. Image B: Tegileridine Dose-Response Curve with similar axes. ED95 is 23.89 ug/kg (95% CI: 23.68-24.12); ED50 is 23.22 ug/kg (95% CI: 22.39-24.19). Horizontal lines at 0.5 and near 0.95; vertical lines at 23.22 and 23.89. Curve near 0 around dose 22, rises steeply around 23-24, plateaus near 0.85 by 24-26. Points at y=0 and y=1 across doses. Image C: Comparative Dose-Response Curves with dose (ug/kg) vs. response probability (0.00-1.00). Fentanyl curve rises around dose 1-2 from 0.2 to 0.55. Tegileridine curve near 0 until 22-23, rises steeply around 23-24, plateaus near 0.85 by 25-26. Legend lists Fentanyl and Tegileridine.Three line graphs showing dose-response curves for fentanyl, tegileridine and a comparative plot. Dose-response curves of tegileridine and fentanyl for inhibiting cervical dilation-induced motor responses. (A) Fentanyl dose-response curve. (B) Tegileridine dose-response curve. (C) Comparative dose-response curves of both drugs. The bootstrap‑based ED50 for tegileridine was 23.22 μg/kg (95% CI: 22.39‑24.19); for fentanyl, 1.51 μg/kg (95% CI: 0.97‑4.07). (For the fentanyl group, a truncated bootstrap retaining only positive estimates was used to avoid negative lower bounds due to small sample size; this does not affect the primary bootstrap inference). To quantify the uncertainty of the potency ratio (tegileridine:fentanyl based on ED50), we calculated the ratio range from the bootstrap confidence limits: lower bound = tegileridine ED50 lower CI/fentanyl ED50 upper CI; upper bound = tegileridine ED50 upper CI/fentanyl ED50 lower CI. The resulting uncertainty range was approximately 1:5.5 to 1:24.9 (based on point estimates: 1:15). This wide range reflects the imprecision of the fentanyl ED50 estimate and underscores the exploratory nature of the finding. As secondary, model‑based estimates, the ED95 values (from probit regression) were 23.89 μg/kg (95% CI: 23.68‑24.12) for tegileridine and 1.97 μg/kg (95% CI: 1.87‑2.07) for fentanyl. These ED95 estimates should be interpreted with caution, as the Dixon up‑and‑down design is optimized for ED50 and provides less precise estimates for upper‑tail quantiles. Forest plots provide a visual comparison of the estimated ED50 and ED95 values with their confidence intervals (Figure 5). Within the dose range determined by the sequential trial, the overall anesthesia success rate remained consistently high in both groups (Figure 6). Figure 5Forest plot of ED50 and ED95 estimates with 95% confidence intervals for tegileridine and fentanyl.A forest plot illustrates ED50 and ED95 estimates with 95% confidence intervals for Tegileridine and Fentanyl. The x-axis, labeled Effective Dose (µg/kg), ranges from 0 to 25. The y-axis lists Tegileridine ED95, Tegileridine ED50, Fentanyl ED95 and Fentanyl ED50. Tegileridine ED95 is near 24 µg/kg with a short confidence interval. Tegileridine ED50 is near 23 µg/kg with a longer interval from about 22 to 24 µg/kg. Fentanyl ED95 is near 2 µg/kg with a very short interval. Fentanyl ED50 is near 1.5 µg/kg with a longer interval from about 1 to 4 µg/kg. A legend clarifies the parameters and drugs.A forest plot comparing ED50 and ED95 estimates for tegileridine and fentanyl, with separated dose levels. Figure 6Comparison of overall response rates (absence of movement during cervical dilation) between the two treatment groups. Data are presented as percentage with actual responder numbers indicated.A bar graph titled Comparative Response Rates. The horizontal axis label is Treatment group, with two categories: Fentanyl and Tegileridine. The vertical axis label is Response Rate left parenthesis percent right parenthesis, ranging from 0 percent to 60 percent with tick marks at 0, 20, 40 and 60. Two vertical bars are shown. Fentanyl has a labeled value of 48.1 percent with responder count left parenthesis 13 slash 27 right parenthesis. Tegileridine has a labeled value of 44.4 percent with responder count left parenthesis 12 slash 27 right parenthesis.A bar graph showing comparative response rates between Fentanyl and Tegileridine. Forest plot of ED50 and ED95 estimates with 95% confidence intervals for tegileridine and fentanyl. Comparison of overall response rates (absence of movement during cervical dilation) between the two treatment groups. Data are presented as percentage with actual responder numbers indicated. Safety Profile Exploratory safety observations are presented in Table 4 and Table 5. The incidence of respiratory depression/hypoxemia was numerically lower in the tegileridine group (14.8% vs. 59.3%). However, these findings are exploratory, derived from a small open‑label sequential dose‑finding study with different administration schedules, rescue medication, and propofol titration. They should not be interpreted as evidence of superior safety at equivalent analgesic levels.Table 4Comparison of Intraoperative Adverse Events Between the Two GroupsAdverse EventTegileridine (n = 27)Fentanyl (n = 27)Total (n = 54)P-valueBradycardia2 (7.4%)5 (18.5%)7 (13.0%)0.422Hypotension13 (48.1%)13 (48.1%)26 (48.1%)1.000Respiratory depression/Hypoxemia4 (14.8%)16 (59.3%)20 (37.0%)0.001At least one adverse even14 (51.9%)23 (85.2%)37 (68.5%)0.017Notes: Data are presented as n (%). Bradycardia was defined as heart rate <60 beats/min; hypotension as systolic blood pressure <90 mmHg or >20% decrease in mean arterial pressure requiring intervention; respiratory depression/hypoxemia as pulse oxygen saturation <92% or respiratory rate <8 breaths/min. P-values were calculated using Fisher’s exact test for between-group comparisons. Bold P-values indicate P < 0.05. P‑values are provided for exploratory purposes only; no adjustment for multiplicity was made, and they should not be interpreted as confirmatory evidence of a difference. Table 5Risk Analysis of Adverse Events: Tegileridine versus FentanylAdverse EventRelative Risk (RR, 95% CI)Risk Difference (RD, 95% CI, %)NNTBradycardia0.40 (0.09–1.86)–11.1 (–27.9 to 5.7)9Hypotension1.00 (0.60–1.67)0.0 (–25.5 to 25.5)–Respiratory depression/Hypoxemia0.25 (0.10–0.65)–44.4 (–66.3 to –22.6)3At least one adverse event0.61 (0.41–0.91)–33.3 (–57.2 to –9.4)3Notes: RR <1 indicates a lower risk in the tegileridine group; RD <0 indicates a lower incidence in the tegileridine group; NNT represents the number of patients needed to be treated with tegileridine instead of fentanyl to prevent one adverse event. Bold values indicate that the 95% CI does not cross 1 for RR, or 0 for RD. These exploratory estimates require confirmation in larger trials. Table 4Comparison of Intraoperative Adverse Events Between the Two GroupsAdverse EventTegileridine (n = 27)Fentanyl (n = 27)Total (n = 54)P-valueBradycardia2 (7.4%)5 (18.5%)7 (13.0%)0.422Hypotension13 (48.1%)13 (48.1%)26 (48.1%)1.000Respiratory depression/Hypoxemia4 (14.8%)16 (59.3%)20 (37.0%)0.001At least one adverse even14 (51.9%)23 (85.2%)37 (68.5%)0.017Notes: Data are presented as n (%). Bradycardia was defined as heart rate <60 beats/min; hypotension as systolic blood pressure <90 mmHg or >20% decrease in mean arterial pressure requiring intervention; respiratory depression/hypoxemia as pulse oxygen saturation <92% or respiratory rate <8 breaths/min. P-values were calculated using Fisher’s exact test for between-group comparisons. Bold P-values indicate P < 0.05. P‑values are provided for exploratory purposes only; no adjustment for multiplicity was made, and they should not be interpreted as confirmatory evidence of a difference. Comparison of Intraoperative Adverse Events Between the Two Groups Notes: Data are presented as n (%). Bradycardia was defined as heart rate <60 beats/min; hypotension as systolic blood pressure <90 mmHg or >20% decrease in mean arterial pressure requiring intervention; respiratory depression/hypoxemia as pulse oxygen saturation <92% or respiratory rate <8 breaths/min. P-values were calculated using Fisher’s exact test for between-group comparisons. Bold P-values indicate P < 0.05. P‑values are provided for exploratory purposes only; no adjustment for multiplicity was made, and they should not be interpreted as confirmatory evidence of a difference. Table 5Risk Analysis of Adverse Events: Tegileridine versus FentanylAdverse EventRelative Risk (RR, 95% CI)Risk Difference (RD, 95% CI, %)NNTBradycardia0.40 (0.09–1.86)–11.1 (–27.9 to 5.7)9Hypotension1.00 (0.60–1.67)0.0 (–25.5 to 25.5)–Respiratory depression/Hypoxemia0.25 (0.10–0.65)–44.4 (–66.3 to –22.6)3At least one adverse event0.61 (0.41–0.91)–33.3 (–57.2 to –9.4)3Notes: RR <1 indicates a lower risk in the tegileridine group; RD <0 indicates a lower incidence in the tegileridine group; NNT represents the number of patients needed to be treated with tegileridine instead of fentanyl to prevent one adverse event. Bold values indicate that the 95% CI does not cross 1 for RR, or 0 for RD. These exploratory estimates require confirmation in larger trials. Risk Analysis of Adverse Events: Tegileridine versus Fentanyl Notes: RR <1 indicates a lower risk in the tegileridine group; RD <0 indicates a lower incidence in the tegileridine group; NNT represents the number of patients needed to be treated with tegileridine instead of fentanyl to prevent one adverse event. Bold values indicate that the 95% CI does not cross 1 for RR, or 0 for RD. These exploratory estimates require confirmation in larger trials. Further risk analysis provides exploratory estimates for tegileridine compared with fentanyl (Table 5). Compared with fentanyl, tegileridine was associated with a relative risk of respiratory depression/hypoxemia of 0.25 (95% CI: 0.10–0.65). The NNT was 3, meaning that replacing fentanyl with tegileridine in three patients would be associated with the prevention of one respiratory depression/hypoxia event in this study. For the composite endpoint of “at least one adverse event”, the risk of occurrence was also lower (RR = 0.61, 95% CI: 0.41–0.91; NNT = 3), although the wide confidence interval indicates uncertainty. No statistically significant difference was observed between the two groups in terms of the risk of hypotension or bradycardia. These exploratory risk estimates should be confirmed in larger trials. Postoperative Nausea and Vomiting (PONV) The incidence of PONV was low in both groups. At 30 minutes postoperatively, nausea/vomiting occurred in 2 patients (7.4%) in the tegileridine group and 3 patients (11.1%) in the fentanyl group (P = 1.000). At 12 hours postoperatively, PONV occurred in 1 patient (3.7%) in each group (P = 1.000). Discussion Key Findings This exploratory dose‑finding study estimated the ED50 of tegileridine and fentanyl for suppressing body movement during hysteroscopic surgery using the Dixon sequential method. Consistent with the design’s strengths, ED50 was the primary endpoint; ED95 estimates were derived secondarily using probit regression and are subject to greater uncertainty. The sequential design is efficient for initial exploration of novel agents.21,22 Clinical Significance and Efficacy Comparison The ED50‑based potency ratio suggests that tegileridine is approximately 15‑fold less potent than fentanyl (point estimate), with a wide uncertainty range of approximately 1:5.5 to 1:24.9 derived from the bootstrap confidence limits. This wide range reflects the imprecision of the fentanyl ED50 estimate (95% CI spanning 0.97‑4.07 μg/kg) and underscores the exploratory nature of the finding. Any dose conversion guidance would require confirmation in a larger, adequately powered trial. Notably, despite requiring a higher absolute dose, tegileridine—as a highly selective G protein‑biased agonist—theoretically possesses a distinct adverse reaction profile compared to conventional agents.;23 however, this remains to be verified in dedicated safety studies. Safety Analysis The safety data are exploratory and hypothesis‑generating only. In this small study, the numerical incidence of respiratory depression was lower in the tegileridine group (14.8% vs. 59.3%). However, these findings cannot be interpreted as proof of superior safety because of the open‑label design, differing infusion durations, use of rescue medication, and propofol titration. The observed difference may be influenced by multiple factors, including the higher propofol consumption in the fentanyl group (though not statistically significant, P = 0.058). Therefore, no conclusion about comparative safety superiority can be drawn; further adequately powered blinded trials are needed. Methodological Considerations This exploratory dose‑finding study employed the modified Dixon sequential method, which is well‑suited for efficiently approximating the ED50 with a modest sample size.21 Conducting the study at a single center helped maintain consistency in anesthesia administration and outcome assessment, reducing operational variability. For statistical inference, we prespecified percentile bootstrap resampling (2,000 replicates) as the primary framework for estimating ED50 confidence intervals.24 Probit regression and four‑parameter logistic modelling were used as complementary sensitivity analyses; they gave consistent results and did not alter the primary conclusions. For the fentanyl group, a truncated bootstrap (retaining only positive estimates) was applied to avoid negative lower bounds due to small sample size; this technical adjustment does not affect the primary bootstrap‑based inference. Simulation studies have shown that the traditional up‑and‑down method may underestimate high‑percentile effective doses (eg., ED95) and is less efficient than modern adaptive designs such as continual reassessment.22 Consistent with this, our ED95 estimates are model‑based and presented as secondary findings. The wide confidence interval for fentanyl ED50 (0.97–4.07 μg/kg) reflects the challenge of precise estimation in small‑sample sequential designs. Nonetheless, the classical Dixon method remains a reasonable choice for initial ED50 estimation in early‑phase exploratory studies because of its operational simplicity and safety.21 Limitations and Future Directions As a preliminary dose-finding study, this research has several limitations. Firstly, the open-label design due to different administration routes (10-min infusion for tegileridine vs. 90-sec bolus with a 4-min wait for fentanyl) may have introduced performance bias. Although outcome assessors were blinded, the unblinded administration of drugs is a limitation, and the different pharmacokinetic profiles may not ensure perfectly equivalent analgesic conditions at the time of cervical dilation, potentially affecting the fairness of comparison. Secondly, the small sample size and sequential design were focused on ED50 estimation, and ED95 was estimated secondarily, which meant that the complete pharmacodynamic profile at both ends of the dose-response curve was not fully characterized. The unexpectedly narrow confidence interval for tegileridine ED95 (23.68–24.12 μg/kg) suggests an unusually steep dose-response curve in this cohort, which warrants further investigation in larger populations. Thirdly, the confidence intervals for the ED50 and ED95 were calculated using bootstrap and sequential methods, respectively, with no unified methodology. The wide confidence interval for fentanyl ED50 (0.97–4.07 μg/kg) reflects the challenge of precisely estimating parameters in the steep middle segment of the dose-response curve in small-sample sequential designs. Fourthly, the high rate of respiratory depression (59.3%) in the fentanyl group is notable and may be related to our sensitive definition (SpO2 < 92% or RR < 8 breaths/min), as well as the well-established respiratory depressant effect of fentanyl itself. Although propofol consumption was similar between groups (P = 0.058), a synergistic effect cannot be entirely excluded. Fifthly, this study primarily focused on intraoperative efficacy and early respiratory safety. Although we collected data on postoperative nausea and vomiting at 30 minutes and 12 hours, the follow-up duration was limited, and other opioid-related side effects such as pruritus and long-term analgesic effects require further investigation. Finally, as a single-center study, the generalizability of our findings to other populations and surgical settings needs to be confirmed in future multicenter trials. Despite these limitations, the core findings of this study provide useful preliminary data. In this exploratory study, we estimated the ED50 of tegileridine fumarate (23.22 μg/kg) and, secondarily, the ED95 (23.89 μg/kg). The ED50-based potency ratio showed wide uncertainty (see Section 3.3). At equivalent analgesic levels, tegileridine was observed to be associated with a numerically lower incidence of respiratory depression compared with fentanyl (14.8% vs. 59.3%, RR = 0.25), a finding consistent with the pharmacological profile of G protein-biased μ-opioid receptor agonists but requiring confirmation in blinded trials.15–17 These findings offer initial dose and safety data for tegileridine in this surgical setting and support further investigation of its potential role in reducing opioid-related adverse events. Conclusions This exploratory dose‑finding study estimated the ED50 of tegileridine (23.22 μg/kg, bootstrap 95% CI 22.39‑24.19) and fentanyl (1.51 μg/kg, 95% CI 0.97‑4.07) for hysteroscopic surgery. The ED50‑based potency ratio (≈1:15) has wide uncertainty (1:5.5‑24.9). Secondary ED95 estimates were obtained from probit regression but must be interpreted cautiously. Exploratory safety observations showed a numerically lower incidence of respiratory depression with tegileridine; however, these findings are not confirmatory. All results require validation in adequately powered, blinded, multicenter trials before clinical application.