Ultrasound-guided lateral pterygoid muscle block as an adjunct to multimodal analgesia for impacted third molar extraction: a two-center retrospective exploratory study Authors: Keisuke Nakazawa, Shun Hoshino, Ryota Tsukui, Motohiro Muranaka, Yuto Oshizaka, Yoshio Ohyama, Junko Tamari, Noriya Hirose, Hisataka Kitano, Takahiro Suzuki Published: 2026 ABSTRACT AbstractBackgroundUltrasound-guided lateral pterygoid muscle (LPM) block, a single-intramuscular-injection technique originally described for sagittal split ramus osteotomy, may offer a simplified regional analgesic option for dental surgery. However, its application as an adjunct to wound infiltration in impacted third molar extraction has not been reported.MethodsThis retrospective two-center exploratory study included 48 consecutive patients who underwent impacted third molar extraction under general anesthesia between September and November 2025. All patients received standardized intraoperative multimodal analgesia (intravenous acetaminophen and dexamethasone) and wound local infiltration; the LPM group (n = 27) additionally received an ultrasound-guided LPM block with 10 ml of 0.25% levobupivacaine, whereas the NE group (n = 21) received wound infiltration alone. The primary outcome was the Numeric Rating Scale (NRS) score at rest 6 h postoperatively. Institutional and extraction-pattern imbalances were addressed using adjusted ordinal logistic regression sensitivity analyses.ResultsThe median NRS score at rest 6 h postoperatively was lower in the LPM group than in the NE group (0 [0, 2] vs. 3 [1.5, 4.0], P < 0.001). The institution- and extraction-pattern-adjusted ordinal logistic regression preserved the direction of the association (adjusted odds ratio 0.14, 95% confidence interval 0.04–0.48, P = 0.002). The recorded total volume of wound local anesthetic was similar between the groups (LPM 9.0 ml vs. NE 8.4 ml, P = 0.729). Lower NRS scores were also observed at 2 h and 21:00; differences resolved by the following morning. The median time to first rescue analgesia was 10 h in the LPM group versus 7 h in the NE group (log-rank P = 0.294). No complications directly attributable to the LPM block were documented during the recorded early postoperative observation window.ConclusionAn adjunctive ultrasound-guided LPM block was associated with lower recorded early postoperative pain scores after impacted third molar extraction. These hypothesis-generating findings require confirmation in the ongoing ALPINE randomized controlled trial (jRCT1031250592).Keywords: Anesthesia, Dental; Molar, Third; Nerve Block; Pain, Postoperative; Pterygoid Muscles; Ultrasonography, Interventional INTRODUCTION Impacted third molar extraction is one of the most commonly performed oral surgical procedures globally [1,2,3]. Suboptimal postoperative pain management contributes to patient discomfort, delayed recovery, and increased demand for rescue analgesics [4]. Traditional approaches to postoperative analgesia for dental surgery rely primarily on systemic analgesics and local wound infiltration. While the intraoral inferior alveolar nerve block has been the standard regional anesthesia technique for decades, it is associated with variable success rates, risks of intravascular injection, and limited applicability in patients with restricted mouth opening (trismus) [5,6,7]. Ultrasound-guided inferior alveolar nerve block improves precision by targeting the pterygomandibular space under direct visualization [8,9], but it still requires accurate needle placement within a narrow anatomical space containing major vasculature, leading to the proposal of extraoral approaches [10]. The lateral pterygoid muscle (LPM) lies in close anatomical proximity to the branches of the trigeminal nerve within the infratemporal fossa and pterygomandibular space [8,11]. We previously reported that ultrasound-guided LPM block—introduced by our group as a single intramuscular injection for perioperative analgesia after sagittal split ramus osteotomy—was technically feasible in a small case series [12]. We use the term “LPM block” as a descriptive shorthand consistent with our previous report [12]. Recent evidence demonstrates that ultrasound-guided extraoral mandibular nerve blocks can provide effective postoperative analgesia in third molar extraction [13], supporting the potential utility of our technique. The aim of this retrospective comparative exploratory study was to evaluate the early postoperative analgesic association of ultrasound-guided LPM block administered as an adjunct to standardized multimodal analgesia and wound local infiltration, compared with wound infiltration alone, to inform the design of an ongoing randomized controlled trial (ALPINE; jRCT1031250592). METHODS 1. Study design and setting This retrospective comparative exploratory study was conducted at Nihon University Itabashi Hospital, Tokyo, Japan, and Shizuoka City Shizuoka Hospital, Shizuoka, Japan. The study protocol was approved by the Institutional Review Boards of Nihon University Itabashi Hospital (RK-251209-5) and Shizuoka City Shizuoka Hospital (25-133). The requirement for written informed consent was waived due to the retrospective design; an opt-out disclosure was provided on institutional websites. The study was conducted in accordance with the Declaration of Helsinki and is reported in accordance with the STROBE statement. Key anesthetic and postoperative analgesic protocol elements were shared across both institutions in preparation for the planned ALPINE prospective randomized controlled trial (jRCT1031250592); shared elements and remaining workflow differences are summarized in Supplementary Table S1. 2. Participants During the study period (September to November 2025), 51 patients underwent screening for inclusion: 31 at Shizuoka City Shizuoka Hospital and 20 at Nihon University Itabashi Hospital. After exclusion of three patients, 48 patients were included in the final analysis. Inclusion criteria were age ≥ 18 years, American Society of Anesthesiologists physical status (ASA-PS) I or II, and unilateral or bilateral third molar extraction. Exclusion criteria were chronic pain conditions, regular analgesic use, psychiatric disorders precluding accurate pain assessment, and inadequate postoperative pain records. Patients who received ultrasound-guided LPM block in addition to wound local anesthesia were assigned to the LPM group; those who received wound local anesthesia alone served as the control (NE) group. 3. Intraoperative anesthesia protocol Intraoperative anesthesia was standardized across both institutions. General anesthesia was induced with propofol, remifentanil, and rocuronium, and the airway was secured via nasotracheal intubation. Immediately after induction, intravenous dexamethasone 6.6 mg was administered to all patients. Anesthesia was maintained with sevoflurane or desflurane and a titrated remifentanil infusion according to clinical requirements. Oral surgeons performed standard wound infiltration at the start of the procedure using 0.5% lidocaine with 1:100,000 epinephrine. At least 15 min before the end of surgery, intravenous acetaminophen was administered at a uniform weight-based dose (1,000 mg for ≥ 50 kg; 15 mg/kg for < 50 kg). Ondansetron 4 mg was administered intravenously for prophylaxis of postoperative nausea and vomiting (PONV). No long-acting opioids were administered intraoperatively. 4. LPM block technique In the LPM group, the ultrasound-guided LPM block was performed after anesthesia induction and before surgery, following the technique described by Nakazawa et al. [12]. With the patient supine, a high-frequency linear transducer was placed in the infratemporal fossa to visualize the LPM. The ultrasound system differed by center: a Venue Go system with an L4-20t-RS linear transducer (GE HealthCare, Chicago, IL, USA) at Nihon University Itabashi Hospital, and a SONIMAGE MX1 system with an L11-3 linear transducer (Konica Minolta, Inc., Tokyo, Japan) at Shizuoka City Shizuoka Hospital. Under real-time guidance, a 22-gauge needle was advanced into the LPM belly using an out-of-plane technique. Following negative aspiration, 10 ml of 0.25% levobupivacaine was injected as a single bolus. For bilateral extractions, the block was performed bilaterally (total dose 50 mg). At each institution, LPM blocks were performed by a designated lead anesthesiologist holding the ESRA European Diploma of Regional Anaesthesia (ESRA-DRA), or by other anesthesiologists under direct supervision. Before injection, color Doppler ultrasonography was performed in every case to identify the maxillary artery and confirm the absence of clinically relevant arterial blood flow along the planned needle trajectory. Negative aspiration was performed as an additional precaution. Representative procedural ultrasound images were not systematically archived; readers are referred to Nakazawa et al. [12] for technical details. 5. Outcomes and data collection The primary outcome was the NRS score (0 = no pain, 10 = worst imaginable pain) at rest 6 h postoperatively, a standardized clinical time point coinciding with the routine postoperative ward round at both institutions. Secondary outcomes included NRS scores at rest and during mouth opening at extubation, 2 h, 21:00, and the following morning at 07:00 and 10:00; the proportion of patients with moderate-to-severe pain (defined analytically as an NRS score > 4); the postoperative rescue-analgesia-free interval (Kaplan–Meier from end of surgery); and complications documented during the recorded early postoperative period. Postoperative NRS scores were recorded by ward nursing staff during routine clinical care using identical standardized observation sheets at both institutions, from emergence until 10:00 on the first postoperative morning. Planned postoperative length of stay differed between institutions (overnight admission at Shizuoka City Shizuoka Hospital and two-night admission at Nihon University Itabashi Hospital); however, the observation window for all study outcomes—including adverse-event documentation—was uniformly restricted to 10:00 on the first postoperative morning at both institutions, and adverse-event ascertainment practices during this window did not differ between institutions despite the difference in planned length of stay. As this was a retrospective study, formal assessor blinding to the anesthetic technique was not implemented or verifiable. 6. Postoperative analgesia Postoperative analgesia followed a uniform, opioid-free protocol. Scheduled oral acetaminophen 400 mg was administered every 8 h starting on the evening of surgery. First-line rescue was oral loxoprofen 60 mg (minimum 6-h interval); intravenous flurbiprofen axetil 50 mg was used when oral intake was not tolerated. Moderate-to-severe pain was defined analytically as an NRS score > 4. Rescue analgesia was administered according to patient-reported pain and clinical judgment; an NRS score ≥ 4 served as a clinical guide rather than a mandatory treatment trigger. 7. Statistical analysis Continuous variables are presented as mean ± SD or median (first quartile, third quartile; Q1, Q3) based on normality (Shapiro–Wilk test) and compared using the Welch t-test or Mann–Whitney U test. Categorical variables are presented as n (%) and compared using Fisher’s exact test. The postoperative rescue-analgesia-free interval was evaluated using the Kaplan–Meier method from the end of surgery and compared via the log-rank test, with administrative censoring at 10:00 on the first postoperative morning. The time-to-event analysis is interpreted descriptively and does not estimate the pharmacodynamic duration of either intervention. The primary analysis compared the 6-h NRS score at rest using the Mann–Whitney U test. Sensitivity analyses used ordinal logistic regression: Adjusted Model 1 included treatment group and institution; Adjusted Model 2 additionally included extraction pattern (mandibular/maxillary-only vs. maxillomandibular) and surgical duration (Supplementary Table S2). Stratified Mann–Whitney U tests were performed within each institution. Subgroup analyses characterized outcomes by extraction pattern (Supplementary Tables S3 and S4). A post hoc sensitivity analysis re-estimated the primary outcome after excluding the first 17 pilot cases used during protocol development. Multiplicity adjustments were not applied to secondary analyses; these are interpreted as exploratory. Statistical analyses were performed by Satista Co., Ltd. (Uji, Kyoto, Japan) using IBM SPSS Statistics version 22.0 (IBM Corp., Armonk, NY, USA). A two-sided P value < 0.05 was considered statistically significant. RESULTS 1. Patient characteristics Of 51 screened patients, 3 were excluded: 2 from the LPM group (bipolar disorder under treatment; incomplete pain records) and 1 from the NE group (incomplete pain records). The final analysis included 48 patients (LPM n = 27; NE n = 21). The 17 pilot cases were retained in the primary analysis as they fell within the predefined study period. The patient flow is shown in Fig. 1. This diagram illustrates the patient flow. Patient demographics were comparable between groups for age, sex, height, and weight (Table 1). Institutional distribution differed significantly (P = 0.003), and the proportion of maxillomandibular extractions tended to be lower in the LPM group (44.4% vs. 71.4%, P = 0.082). Surgical time was comparable (P = 0.454), and the recorded total volume of wound local anesthetic was similar between groups (LPM 9.0 ml vs. NE 8.4 ml, P = 0.729). Table 1Patient demographics and baseline characteristicsVariableLPM group (n = 27)NE group (n = 21)P valueAge, yr39.4 ± 13.642.0 ± 18.20.602Female sex, n (%)22 (81.5)15 (71.4)0.498Height, cm160.3 ± 7.5161.0 ± 8.50.766Weight, kg58.2 ± 12.859.8 ± 15.00.699Institution, n (%)0.003Nihon University6 (22.2)14 (66.7)Shizuoka City21 (77.8)7 (33.3)ASA-PS*, n (%)1.000I26 (96.3)20 (95.2)II1 (3.7)1 (4.8)Extraction pattern, n (%)0.082Maxillomandibular12 (44.4)15 (71.4)Mandibular only14 (51.9)5 (23.8)Maxillary only1 (3.7)1 (4.8)Surgical time, min48 (38, 70)59 (37, 87)0.454Remifentanil, mg1.4 (1.3, 2.0)1.6 (1.0, 2.2)0.699Wound local anesthesia, ml†9.0 (5.0, 10.0)8.4 (6.6, 10.0)0.729Values are mean ± standard deviation, median (Q1, Q3), or n (%), as appropriate. Continuous variables with approximately normal distributions (age, height, weight) are presented as mean ± SD and were compared using the Welch t-test. Surgical time, remifentanil dose, and wound local anesthetic volume are non-normally distributed and are presented as median (Q1, Q3); they were compared using the Mann–Whitney U test. Categorical variables were compared using Fisher’s exact test. Institution was placed immediately after demographic variables to highlight the significant baseline imbalance, which is addressed in the Limitations and in sensitivity analyses (Supplementary Table S2).*ASA-PS, American Society of Anesthesiologists physical status classification. †Recorded total volume of wound local anesthetic (0.5% lidocaine with 1:100,000 epinephrine) used in the surgical field. Tooth-specific or anatomical-site-specific infiltration volumes were not recorded.LPM, lateral pterygoid muscle block; NE, no extra regional block (wound local anesthesia alone); Q1, first quartile; Q3, third quartile; SD, standard deviation Patient demographics and baseline characteristics Values are mean ± standard deviation, median (Q1, Q3), or n (%), as appropriate. Continuous variables with approximately normal distributions (age, height, weight) are presented as mean ± SD and were compared using the Welch t-test. Surgical time, remifentanil dose, and wound local anesthetic volume are non-normally distributed and are presented as median (Q1, Q3); they were compared using the Mann–Whitney U test. Categorical variables were compared using Fisher’s exact test. Institution was placed immediately after demographic variables to highlight the significant baseline imbalance, which is addressed in the Limitations and in sensitivity analyses (Supplementary Table S2). *ASA-PS, American Society of Anesthesiologists physical status classification. †Recorded total volume of wound local anesthetic (0.5% lidocaine with 1:100,000 epinephrine) used in the surgical field. Tooth-specific or anatomical-site-specific infiltration volumes were not recorded. LPM, lateral pterygoid muscle block; NE, no extra regional block (wound local anesthesia alone); Q1, first quartile; Q3, third quartile; SD, standard deviation 2. Primary outcome The median NRS score at rest 6 h postoperatively was lower in the LPM group than in the NE group (0 [0, 2] vs. 3 [1.5, 4.0], P < 0.001; Table 2). The institution-adjusted ordinal logistic regression yielded an adjusted odds ratio of 0.13 (95% CI 0.04–0.46, P = 0.001) for the LPM group (Supplementary Table S2, Adjusted Model 1). In Adjusted Model 2, controlling simultaneously for institution, extraction pattern, and surgical duration, the direction of the association remained unchanged (adjusted odds ratio 0.14, 95% CI 0.04–0.48, P = 0.002); neither extraction pattern (P = 0.844) nor surgical duration (P = 0.699) was associated with the primary outcome. Stratified analysis confirmed lower NRS scores in the LPM group within both institutions (Nihon University: P = 0.040; Shizuoka City: P = 0.013). In the post hoc sensitivity analysis excluding the 17 pilot cases, the between-group difference remained significant (LPM 0 [0, 2] vs. NE 3 [2, 4.5], P < 0.001, n = 31). Given the floor effect in the LPM group, the adjusted point estimates should be interpreted as supportive sensitivity evidence rather than as stable causal estimates. Table 2Numeric Rating Scale pain scores at restTime pointLPM group (n = 27)NE group (n = 21)P valueExtubation0 (0, 0)0 (0, 0)0.2572 h0 (0, 0)3 (0, 5)< 0.001 6 h (primary) 0 (0, 2) 3 (1.5, 4.0) < 0.001 21:000 (0, 2)3 (1.5, 5.0)< 0.00107:00 (next morning)1 (0, 4)3 (0.5, 4.0)0.21310:00 (next morning)1 (0, 3)3 (1.5, 3.0)0.074Values are median (Q1, Q3). P values were calculated using the Mann–Whitney U test.The primary outcome (6 h) is shown in bold. h, hour; LPM, lateral pterygoid muscle block; NE, no extra regional block; NRS, Numeric Rating Scale (0 = no pain, 10 = worst imaginable pain); Q1, first quartile; Q3, third quartile. Numeric Rating Scale pain scores at rest Values are median (Q1, Q3). P values were calculated using the Mann–Whitney U test. The primary outcome (6 h) is shown in bold. h, hour; LPM, lateral pterygoid muscle block; NE, no extra regional block; NRS, Numeric Rating Scale (0 = no pain, 10 = worst imaginable pain); Q1, first quartile; Q3, third quartile. 3. Secondary outcomes NRS scores at rest were lower in the LPM group at 2 h postoperatively (0 [0, 0] vs. 3 [0, 5], P < 0.001) and at 21:00 (0 [0, 2] vs. 3 [1.5, 5], P < 0.001). No differences were observed at extubation, 07:00, or 10:00 the following morning (Table 2). NRS scores during mouth opening followed a similar pattern, with lower scores in the LPM group at 2 h, 6 h, and 21:00 (all P < 0.001; Table 3). Table 3Numeric Rating Scale pain scores during mouth openingTime pointLPM group (n = 27)NE group (n = 21)P valueExtubation0 (0, 0)0 (0, 0)0.2572 h0 (0, 0)3 (0.5, 5)< 0.0016 h0 (0, 3)3 (2, 5)< 0.00121:001 (0, 3)4 (3, 6)< 0.00107:00 (next morning)2 (0, 4)3 (1, 5)0.37110:00 (next morning)2 (0, 4)3 (2, 4)0.129Values are median (Q1, Q3). P values were calculated using the Mann–Whitney U test.h, hour; LPM, lateral pterygoid muscle block; NE, no extra regional block; NRS, Numeric Rating Scale (0 = no pain, 10 = worst imaginable pain); Q1, first quartile; Q3, third quartile. Numeric Rating Scale pain scores during mouth opening Values are median (Q1, Q3). P values were calculated using the Mann–Whitney U test. h, hour; LPM, lateral pterygoid muscle block; NE, no extra regional block; NRS, Numeric Rating Scale (0 = no pain, 10 = worst imaginable pain); Q1, first quartile; Q3, third quartile. The proportion of patients with an NRS score > 4 at rest was lower in the LPM group at 2 h (7.4% vs. 38.1%, P = 0.013) and at 21:00 (3.7% vs. 28.6%, P = 0.034). During mouth opening, the LPM group showed lower proportions at 2 h (11.1% vs. 42.9%, P = 0.018), 6 h (3.7% vs. 28.6%, P = 0.034), and 21:00 (3.7% vs. 42.9%, P = 0.001; Table 4). Table 4Proportion of patients with moderate-to-severe pain (NRS > 4)Time pointLPM group, n/N (%)NE group, n/N (%)P valueNRS > 4 at restExtubation0/27 (0)0/21 (0)–2 h2/27 (7.4)8/21 (38.1)0.0136 h1/27 (3.7)3/21 (14.3)0.30621:001/27 (3.7)6/21 (28.6)0.03407:00 (next morning)4/27 (14.8)3/21 (14.3)1.00010:00 (next morning)5/27 (18.5)1/21 (4.8)0.211NRS > 4 during mouth openingExtubation0/27 (0)0/21 (0)–2 h3/27 (11.1)9/21 (42.9)0.0186 h1/27 (3.7)6/21 (28.6)0.03421:001/27 (3.7)9/21 (42.9)0.00107:00 (next morning)6/27 (22.2)6/21 (28.6)0.74110:00 (next morning)6/27 (22.2)3/21 (14.3)0.712Proportions are expressed as number of patients with NRS > 4 / total assessed (%). P values were calculated using Fisher’s exact test.Moderate-to-severe pain was defined analytically as NRS > 4. This analytical definition is distinct from the clinical guidance for rescue analgesia, which was based on patient request and clinical judgment.–, not applicable; h, hour; LPM, lateral pterygoid muscle block; NE, no extra regional block; NRS, Numeric Rating Scale. Proportion of patients with moderate-to-severe pain (NRS > 4) Proportions are expressed as number of patients with NRS > 4 / total assessed (%). P values were calculated using Fisher’s exact test. Moderate-to-severe pain was defined analytically as NRS > 4. This analytical definition is distinct from the clinical guidance for rescue analgesia, which was based on patient request and clinical judgment. –, not applicable; h, hour; LPM, lateral pterygoid muscle block; NE, no extra regional block; NRS, Numeric Rating Scale. The median postoperative rescue-analgesia-free interval from the end of surgery was 10 h in the LPM group versus 7 h in the NE group (log-rank P = 0.294; Fig. 2). The total number of rescue-analgesic administrations during the recorded early postoperative observation period did not differ significantly (LPM 1 [1, 2] vs. NE 2 [1, 2], P = 0.286; Supplementary Table S5). Rescue-analgesia outcomes by extraction pattern are presented in Supplementary Table S4. The Kaplan–Meier curves display the postoperative rescue-analgesia-free interval. The curves show the probability of remaining rescue-analgesia-free after the end of surgery in the lateral pterygoid muscle (LPM) block group and the non-exposed (NE) control group. The solid line represents the LPM group, and the dashed line represents the NE group. Tick marks indicate censored observations. The numbers at risk are shown below the graph. The median postoperative rescue-analgesia-free interval was 10 h in the LPM group and 7 h in the NE group; the difference was not statistically significant by the log-rank test (P = 0.294). LPM, lateral pterygoid muscle block; NE, wound infiltration alone control group. 4. Subgroup analysis Subgroup analyses are presented in Supplementary Table S3. In maxillomandibular extraction (LPM n = 12, NE n = 15), NRS score at rest was lower in the LPM group at 2 h (P = 0.002), 6 h (P = 0.023), and 21:00 (P = 0.022). In mandibular-only extraction (LPM n = 14, NE n = 5), NRS score at rest was lower in the LPM group at 2 h (P = 0.005), 6 h (P < 0.001), and 21:00 (P = 0.006). Given the very small mandibular-only NE subgroup (n = 5), this comparison should be regarded as descriptive only. Within the LPM group, NRS scores did not differ significantly between extraction patterns (6 h: P = 0.255); given small within-group sample sizes, this does not establish equivalence. 5. Complications No complications directly attributable to the LPM block were documented during the recorded early postoperative observation period (until 10:00 on the first postoperative morning). PONV occurred in 0/27 (0%) patients in the LPM group versus 1/21 (4.8%) in the NE group (P = 0.438). No cases of clinically apparent trismus were reported. Complete adverse event data are presented in Supplementary Table S5. Structured surveillance for delayed complications was not performed. DISCUSSION In this two-center retrospective exploratory study, adjunctive ultrasound-guided LPM block was associated with lower NRS scores at rest and during mouth opening at 2 h, 6 h, and 21:00 compared with wound local infiltration alone. Differences resolved by the following morning. The association remained directionally consistent in sensitivity analyses controlling for institution, extraction pattern, and surgical duration (adjusted odds ratio 0.14, P = 0.002). The recorded total volume of wound local anesthetic was similar between groups (P = 0.729), making greater wound local anesthetic exposure in the LPM group an unlikely explanation. No block-related complications were documented during the recorded early postoperative window. 1. Adjunct-to-wound-infiltration clarification As all patients in both groups received wound local anesthesia, the present comparison evaluated the incremental association of adding LPM block to background wound infiltration, rather than comparing mutually exclusive field-infiltration and proximal-block strategies. Peripheral sensory contributions from accessory innervation in the mandibular retromolar area, including the long buccal nerve and communicating sensory branches, may have been at least partly addressed in both groups by wound infiltration. Although anatomical continuity may permit spread of injectate beyond the immediate intramuscular injection site, injectate distribution and sensory coverage were not systematically assessed; the present data cannot establish involvement of the maxillary nerve, the long buccal nerve, or other accessory pathways. 2. Mechanistic considerations and limitations The analgesic mechanism of intramuscular LPM injection remains unestablished. We did not systematically image injectate spread or perform sensory mapping. Cadaveric and technical literature involving the pterygomandibular space [8] or pterygopalatine fossa approaches [11,14] concern target compartments and needle trajectories anatomically distinct from the intramuscular LPM injection used here; their findings cannot be directly extrapolated. As the present study did not systematically assess interincisal opening or masticatory muscle function, we cannot attribute the observed pain-score differences to muscle relaxation or reduced trismus. 3. Relationship to ultrasound-guided inferior alveolar nerve block The two approaches target conceptually distinct anatomical structures. Ultrasound-guided inferior alveolar nerve block deposits local anesthetic into the pterygomandibular space adjacent to the mandibular foramen, whereas LPM block places the injectate within the LPM itself. Although the LPM is a larger sonographic target, it is a deep intramuscular target close to the maxillary artery, and restricted mouth opening or obesity may pose technical challenges for LPM block as well. Any procedural advantage of one approach over the other requires direct prospective comparison. The early postoperative analgesic association observed is consistent with previous reports using long-acting amide local anesthetics for third molar extraction analgesia [15], and a similar analgesic benefit for an extraoral mandibular nerve approach has been reported by an independent group [13]. 4. Rescue-analgesia discordance The difference in NRS scores at scheduled assessments was not accompanied by a statistically significant difference in time to first rescue analgesia or in the number of rescue-analgesic administrations. Multiple factors related to the retrospective design likely contributed: the rescue threshold was not enforced; combination of different-mechanism agents may have decoupled rescue counts from pain intensity; administrative censoring at 10:00 on the first postoperative morning may have limited the ability to evaluate later differences; and statistical power for the time-to-event analysis was limited. The present findings support an association with lower recorded early pain intensity but do not establish a reduced requirement for additional analgesia. 5. Limitations Some limitations warrant consideration. First, the retrospective, non-randomized design with significant institutional imbalance (P = 0.003) and a tendency toward imbalance in extraction pattern (P = 0.082) introduces residual confounding that cannot be fully resolved at this sample size. Institution may have acted as a composite proxy for differences in surgical workflow, operator experience, nursing assessment culture, and discharge workflow; the institution-adjusted estimate should be interpreted as a directional sensitivity indicator. Recorded wound local anesthetic volumes were similar between groups (P = 0.729), making this co-intervention an unlikely explanation. Second, the comparison evaluated an incremental association beyond baseline wound infiltration, not the intrinsic analgesic territory of the LPM block in isolation. Third, the time-to-event analysis began at the end of surgery and represents a postoperative rescue-analgesia-free interval rather than the pharmacodynamic duration of the intervention; exact per-patient injection times were not extracted, precluding an injection-time-anchored re-analysis. Fourth, safety assessment was limited to adverse events documented during the early postoperative window (until 10:00). Delayed complications including persistent sensory disturbance, prolonged trismus, masticatory dysfunction, hematoma, infection, myotoxicity, and rare acute injection-related events such as vascular injury or local anesthetic systemic toxicity were not systematically monitored. Color Doppler ultrasonography and negative aspiration were performed as procedural precautions, but the absence of documented early complications cannot establish a definitive medium-term safety profile. Fifth, the primary outcome exhibited a strong floor effect in the LPM group (median NRS score = 0), which may inflate the ordinal logistic regression point estimate. Sixth, the 17 pilot cases were included in the main analysis; although a post hoc sensitivity analysis excluding these cases yielded consistent results, this approach may introduce optimistic bias. Seventh, multiple comparisons were performed without formal multiplicity adjustment; secondary findings are exploratory. Eighth, generalizability is limited by the specific anesthetic protocol, predominantly ASA I Japanese population, and opioid-free regimen. Procedural feasibility metrics (block success rate, performance time, image acquisition difficulty, sensory block extent) were not collected, and representative intraoperative ultrasound images were not systematically archived. Formal assessor blinding to the anesthetic technique was not implemented or verifiable given the retrospective design. In conclusion, an adjunctive ultrasound-guided LPM block was associated with lower recorded early postoperative pain scores after impacted third molar extraction. Given the non-randomized design, institutional and extraction-pattern imbalances, and limited safety surveillance, these findings should be regarded as hypothesis-generating and require confirmation in the ongoing ALPINE prospective randomized controlled trial (jRCT1031250592).