Ultrasound guidance does not improve local anesthetic distribution in suprazygomatic maxillary nerve blocks in pediatric patients: a clinical, randomized, controlled, observer-blinded, crossover MRI trial Authors: Nergis Nina Suleiman, Ingrid Lien, Mehdi Sadat Akhavi, Davide Impieri, Kyrre Ullensvang, Jon Magnus Moen, Pernille Lindholm, Eldrid Langesæter, Axel R Sauter Published: 2026 ABSTRACT ABSTRACTBackgroundMaxillary nerve blocks are often used as part of a multimodal postoperative pain management regimen for children undergoing cleft palate surgery. It has not been demonstrated yet that ultrasound guidance provides superior local anesthetic distribution and better success rates compared with the landmark-based block technique. The aim of the study was to compare the effect of ultrasound guidance for suprazygomatic maxillary nerve blocks on local anesthetic spread to the pterygopalatine fossa.Methods20 pediatric patients scheduled for cleft palate repair received bilateral suprazygomatic maxillary blocks. Landmark technique was used on one side and ultrasound guidance on the contralateral side according to randomization. After the block procedure, MRI was used to determine the spread of the local anesthetic.ResultsLocal anesthetic spread to the pterygopalatine fossa with contact to the maxillary nerve (defined as successful block) was observed in 13 (65%) of the landmark-based blocks and in 14 (70%) of the ultrasound-guided blocks (p=0.56). No adverse events or complications were observed during or after the block procedures.ConclusionAppropriate pain management for cleft palate repair can be achieved without the availability of imaging guidance. Future studies should focus on improving the success rate of maxillary nerve blocks, both with and without the aid of sonography.Trial registration numberNCT05778903. WHAT IS ALREADY KNOWN ON THIS TOPIC Suprazygomatic maxillary nerve blocks can improve postoperative pain management in children undergoing cleft palate surgery. Suprazygomatic maxillary nerve blocks can improve postoperative pain management in children undergoing cleft palate surgery. Blocks are performed using either landmark techniques or ultrasound guidance. Blocks are performed using either landmark techniques or ultrasound guidance. WHAT THIS STUDY ADDS Local anesthetic spread to the maxillary nerve in the pterygopalatine fossa can be confirmed in two-thirds of the blocks, regardless of the guidance technique. Local anesthetic spread to the maxillary nerve in the pterygopalatine fossa can be confirmed in two-thirds of the blocks, regardless of the guidance technique. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Landmark-guided techniques can be considered a viable option that does not rely on the availability of an ultrasound machine. Landmark-guided techniques can be considered a viable option that does not rely on the availability of an ultrasound machine. Future research should focus on how to improve maxillary nerve blocks with and without ultrasound guidance. Future research should focus on how to improve maxillary nerve blocks with and without ultrasound guidance. Introduction Cleft palate is one of the most prevalent congenital birth defects, with an incidence of 1–25 per 10 000 live births.1 Surgical repair is essential for improving critical functions such as feeding and speech development. Surgery is typically performed before the age of 2 years.2 3 Effective postoperative pain management is crucial for enhancing recovery, as inadequate pain control can lead to agitation and crying, which may result in complications such as wound dehiscence and fistula formation. In contrast, children who receive adequate pain relief are more likely to resume normal feeding and drinking behaviors earlier in the postoperative period. Maxillary nerve blocks are often used as part of a multimodal postoperative pain management regimen for children undergoing cleft palate surgery.49 These blocks are endorsed by the procedure-specific postoperative pain management (PROSPECT) guidelines for perioperative pain management after cleft palate surgery.10 Traditionally, maxillary nerve blocks are performed using landmark-guided techniques, with the suprazygomatic approach being the most commonly used. This approach involves advancing the needle in a safe direction, carefully avoiding critical structures in the cranium and orbit.49 Recently, ultrasound-guided techniques have been proposed as an alternative to landmark-based approaches.5 6 9 11 Ultrasound guidance offers the theoretical advantage of real-time visualization of both needle placement and local anesthetic spread, potentially increasing block accuracy and safety. However, it has not been demonstrated yet that ultrasound guidance provides superior local anesthetic distribution and therefore better success rates compared with the landmark-based block technique. The key determinant of a successful maxillary nerve block is the spread of local anesthetic to the pterygopalatine fossa and maxillary nerve. MRI with a T2-weighted sequence has been used in previous studies to identify local anesthetic distribution after peripheral block procedures.12 13 In our study, we wanted to compare the local anesthetic distribution after a suprazygomatic maxillary nerve block as visualized by MRI using the landmark technique on one side of the patients and under ultrasound guidance on the contralateral side. The primary aim of the study was to evaluate the local anesthetic spread to the pterygopalatine fossa and maxillary nerve (defined as block success) comparing ultrasound guidance with a landmark-guided technique. A secondary aim was to investigate the spread of the injectate to other anatomical structures. We hypothesized that ultrasound guidance would increase the likelihood of successful local anesthetic distribution to the pterygopalatine fossa and maxillary nerve, compared with the traditional landmark-based technique. Methods Ethics This randomized, controlled, observer-blinded, cross-over study was conducted at Oslo University Hospital between June 7, 2023 (first patient enrolled) and June 19, 2024. The study was prospectively registered in ClinicalTrials.gov (NCT05778903; March 20, 2023). Study subjects Twenty pediatric patients scheduled for cleft palate surgery with suprazygomatic maxillary nerve block as part of their perioperative pain management were included in the study. Only non-syndromic children without craniofacial malformations other than cleft palate (and cleft lip), aged 10–24 months, with American Society of Anesthesiologists status 1 or 2 were included. Exclusion criteria were concomitant medical treatments or conditions that interfered with MRI or peripheral nerve block treatment, allergy to bupivacaine, or other local anesthetics. Anesthesia induction and maintenance Patients were premedicated with nasal dexmedetomidine 3 µg/kg. Standard monitoring included non-invasive blood pressure, electrocardiography, and pulse oximetry. General anesthesia was induced with nitrous oxide and sevoflurane. Two intravenous lines were placed after induction. The children received fentanyl (1.5 µg/kg), propofol (3 mg/kg), and mivacurium (0.2 mg/kg) before endotracheal intubation. End-tidal CO2 was monitored during the perioperative period. Anesthesia was maintained with isoflurane during the block procedure, the MRI examination, and the subsequent surgical procedure. A hybrid operating theater with operating facilities and MRI scanner was used for the treatment and examination of the patients. Block procedures During the study, all patients received bilateral suprazygomatic maxillary blocks. The landmark technique was used on one side, and ultrasound guidance was used on the contralateral side. A 50 mm, 25-gage needle (SonoBlock, Facet S tip, Pajunk, Geisingen, Germany) was used for both techniques. All patients were in the supine position. A volume of 0.15 mL/kg of bupivacaine (2.5 mg/mL) with epinephrine (5 µg/kg) was used for the nerve blocks on each side. Aspiration was performed prior to the injection of the local anesthetic. For landmark-guided suprazygomatic nerve blocks, the method described by Chiono et al was followed.5 After skin disinfection, the needle was inserted perpendicular to the skin at the frontozygomatic angle and advanced until the needle tip was in contact with the greater wing of the sphenoid. The needle was then withdrawn a few millimeters and redirected anteriorly and caudally toward the nasolabial fold (figure 1). Landmark-guided suprazygomatic maxillary nerve block. (A) The needle is inserted perpendicular to the skin at the frontozygomatic angle and advanced until the needle tip is in contact with the greater wing of the sphenoid. (B) The needle is then withdrawn a few millimeters and redirected anteriorly and caudally toward the nasolabial fold. A Venue ultrasound system (GE HealthCare, Mexico) with an L4-20T linear array transducer was used for ultrasound-guided maxillary blocks. Sterile preparation included skin disinfection and sterile transducer covers. The transducer was positioned superficially and parallel to the zygomatic arch. By moving the transducer cranially, the acoustic shadow of the zygomatic arch disappears and the infratemporal fossa can be identified between the maxilla and the sphenoid bone. The probe was adjusted by gentle rotation and tilting until the infratemporal fossa deepened into a narrow cleft representing the pterygopalatine fossa. The needle was then inserted cranially to the transducer and advanced out-of-plane at a steep angle, almost parallel to the ultrasound beam, toward the pterygopalatine fossa (figure 2). Suprazygomatic ultrasound visualization and needle placement in the pterygopalatine fossa. (A) Simulation of the ultrasound scan and needle placement using a pediatric skull model in a water tank. (B) Ultrasound image of the bony landmarks and needle using a pediatric skull model in a water tank. The triangles mark the position of the needle. (C) Suprazygomatic ultrasound scan and needle placement in a pediatric patient. The transducer is initially positioned superficially and parallel to the zygomatic arch. By moving the transducer cranially, the infratemporal fossa can be identified. The probe must be adjusted by gentle rotation and tilting until the infratemporal fossa deepens into a narrow gap representing the pterygopalatine fossa. The needle is inserted cranially to the transducer and advanced at a steep angle. (D) Ultrasound image of a suprazygomatic ultrasound scan and needle placement. The red arrow indicates needle insertion. MB, maxillary bone; PPF, pterygopalatine fossa; SB, sphenoid bone; TB, temporal bone; ZA, zygomatic arch. Outcome assessments The primary outcome was fluid accumulation in the pterygopalatine fossa with contact to the maxillary nerve as seen on the MRI scan after performing a suprazygomatic maxillary block (defined as successful block). Visualization of other anatomical areas (infratemporal fossa, orbit, nasal cavity, cranium) was a secondary outcome. Other secondary outcomes were needle depths and needle angles, blood aspiration, bleeding or hematoma related to the block procedures, and the occurrence of side effects and adverse events. Time spent on MRI and morphine consumption in the first 24 hours postoperatively were also recorded. Registration of needle positions Needle angles were measured as proposed by Prigge et al using a digital protractor (Angle Finder—Digital Measure for Android).14 Measurement of superior/inferior needle angle: the axial plane was considered as 0°; if the needle was angled superiorly, this was considered as an increase (+) in the angle, while any inferior angulation of the needle was considered as a decrease (−) (as seen from anterior perspective). Measurement of posterior/anterior needle angle: the coronal plane was considered to be 0°, with any anterior deviation being considered as an increase (+) and any posterior deviation as a decrease (−) in the angle (as seen from superior perspective) (online supplemental file 1). Needle depth was determined by placing a fingertip on the needle shaft at skin level before removing the needle after block procedures. The depth was then measured using a ruler. Magnetic resonance imaging MRI scanning was performed using a 3T Ingenia scanner (Philips Healthcare, The Netherlands) at the level of the skull base through the maxilla. Initially, as part of a pilot study, sequences were acquired both before and after the injection. Based on the experience of the pilot study, it was decided to perform only postinjection MRI to optimize workflow and reduce time in the hybrid operating room, as pre-injection imaging was deemed unnecessary. The post-injection imaging included the following sequences: three-dimensional (3D) T1-weighted imaging (3D BrainVIEW), T2-weighted turbo spin-echo (TSE) imaging generating in-phase, opposed-phase, water-only, and fat-only images (mDixon), and 3D high-resolution T2-weighted TSE imaging (3D DRIVE). Injectate distribution was classified according to its localization and extent within the imaging planes. Perineural spread was defined as direct contact between the injectate and relevant neural structures at the examined levels. Randomization and blinding Patients were allocated to either arm 1 (ultrasound-guided maxillary nerve block on the left side of the patient and landmark-based maxillary nerve block on the right side of the patient) or arm 2 (landmark-based maxillary nerve block on the left side of the patient and ultrasound-guided maxillary nerve block on the right side of the patient) using computerized randomization. Twenty preprinted sheets with the randomly assigned sequences were placed in 20 identical opaque sealed envelopes marked with identification numbers. The envelopes were opened immediately before the block procedure. Irrespective of the assigned block technique, blocks were performed first on the left and then on the right side of the patient. Landmark blocks were performed by three different surgeons (NNS, DI, PL), and ultrasound-guided blocks were performed by three anesthesiologists (IL, KU, ARS). Each surgeon or anesthesiologist performed six or seven of the block procedures. An independent radiologist, blinded to allocation and not present during the block procedures, analyzed the MRI images. Blinding was maintained until the end of data collection. Group allocation was coded during statistical analysis and was not revealed until the first draft of this manuscript was completed. Statistical analysis The predefined main outcome variable in this study was local anesthetic distribution in the pterygopalatine fossa with a paired study design. We wanted to investigate whether ultrasound could improve suprazygomatic maxillary nerve blocks. Since a negative effect of ultrasound visualization was not expected, a one-sided test was considered appropriate. We considered a 25% increase in successful propagation to the PPF to be clinically significant. To detect a 25% increase (from 65% to 90%) with an alpha value of 0.05 and a power (beta) value of 80%, a total of 18 patients was found to be necessary. To account for missing data, 20 patients were included. Statistical analysis was done using Stata/SE V.15.1 (StataCorp). Categorical data were presented as frequencies and percentages, and continuous data were presented as means, SD and ranges. Differences between paired categorical variables were analyzed using McNemar’s test. For continuous data, Student’s t-test or analysis of variance for paired data was used. P values <0.05 were considered statistically significant. There was only one identified primary analysis in this study, and no multiplicity correction was performed. All other analyses were considered supportive or exploratory. Results Twenty pediatric patients scheduled for cleft palate surgery whose legal guardians gave written informed consent were included in the study (figure 3). Seven (35%) of the participants were male and 13 were female (65%). They had a mean age of 12±3.1 months (range 10–23 months) and a mean weight of 9.8±1.4 kg (range 6.9–12.5 kg). Modified Consolidated Standards of Reporting Trials flow diagram. SMNB, suprazygomatic nerve blocks. In all 20 patients, suprazygomatic maxillary nerve blocks could be performed with landmark technique or ultrasound guidance according to randomization and group allocation. The mean needle depth was 42.4±3.9 mm (range 34–50 mm) for landmark-guided blocks and 44.4±4.2 mm (range 38–50 mm) for ultrasound-guided blocks (p=0.03). The mean superior/inferior needle angle with landmark guidance was 37.5°±9.8° (range 16°–55°) and 29.9°±10.5° (range 10.6°–48°) when ultrasound guidance was used (p<0.01). The mean anterior/posterior needle angles were 15.6°±6.6° (range 7°–31°) and 12.2°±6.5° (range −3°–23°) for landmark-guided blocks and ultrasound-guided blocks, respectively (p=0.15). Blood aspiration and subsequent needle redirection were observed twice during landmark-guided maxillary nerve blocks, whereas no blood aspiration was observed during ultrasound-guided blocks (p=0.16). No bleeding or hematoma was observed after either landmark-guided or ultrasound-guided blocks. No other adverse events or complications were observed during or after the block procedures. All 20 pediatric patients underwent examination by MRI immediately after maxillary nerve block. The mean time for the examination was 47±6 min (range 35–60 min). Local anesthetic spread to the pterygopalatine fossa with contact to the maxillary nerve (defined as successful block) was observed in 13 (65%) of the landmark-guided blocks and in 14 (70%) of the ultrasound-guided blocks (p=0.56). Interestingly, 12 (60%) of the patients had bilateral spread to the pterygopalatine fossa (with ultrasound guidance on one side and the landmark technique on the other side), while bilateral lack of local anesthetic spread was observed in five (25%) patients. Figure 4 shows an MRI image with fluid accumulation in the pterygopalatine fossa and spread into the infratemporal fossa. A detailed characterization of fluids visualized by MRI is shown in table 1. In two of the ultrasound-guided blocks, the spread of the injectate could not be visualized at all (neither in the pterygopalatine fossa, the infratemporal fossa, nor in other locations). Table 1Spread of the local anesthetic injectate as visualized by MRILandmark-guided blocks (n=20)Ultrasound-guided blocks (n=20)Injectate within the pterygopalatine fossa in contact with maxillary nerve13 (65%)14 (70%)Injectate adjacent to pterygopalatine fossa15 (75%)14 (70%)Injectate within infratemporal fossa20 (100%)18 (90%)Injectate within orbit0 (0%)0 (0%)Injectate within nasal cavity1 (5%)1 (5%)Injectate within cranium0 (0%)0 (0%)No statistically significant differences were found between ultrasound-guided and landmark-guided block techniques. Spread of the local anesthetic injectate as visualized by MRI No statistically significant differences were found between ultrasound-guided and landmark-guided block techniques. MRI examination after suprazygomatic nerve block (T2-weighted DRIVE sequence, axial plane). (A) MRI image with labeling of relevant anatomical structures. The infratemporal fossa is colored green, the pterygopalatine fossa is colored red. (B) The arrows indicate the spread of the local anesthetic (shown as a white pattern) within the pterygopalatine fossa and the infratemporal fossa. (C) Axillary plane MRI at a slightly more cranial level showing the maxillary nerve with adjacent blood vessel. Both structures have a black appearance. Identification is based on tracing the structures in several planes. Local anesthetic injectate (marked by the arrows) is in contact with the nerve. During the first 24 hours postoperatively, the median morphine consumption (rectal or oral) was 0.51 mg/kg (range 0.30–0.76 mg/kg) (online supplemental file 2). There was no statistically significant difference between patients with successful bilateral block compared with those who only had unilateral fluid accumulation in the pterygopalatine fossa and those who had complete block failure (according to our definition of block success/failure). Discussion In this randomized, controlled, observer-blinded, crossover study, we compared the use of superficial landmarks with ultrasound guidance for suprazygomatic maxillary nerve block in children undergoing cleft palate repair. A successful maxillary nerve block was defined by MRI visualization as fluid accumulation in the pterygopalatine fossa with contact to the maxillary nerve. Similar results were found for both methods, with success rates ranging from 65% for landmark guidance to 70% for ultrasound guidance. Interestingly, unsuccessful block was often bilateral (both with the landmark technique and ultrasound guidance). This suggests that block failure may depend less on the method used and more on the anatomy of the individual patient. However, we have not yet been able to identify MRI-based criteria that predict unsuccessful blocks. In two of the ultrasound-guided maxillary nerve blocks, no local anesthetic was detected on MRI. The absence of injectate could be explained by intravascular injection. Several previous studies have investigated the analgesic effect of maxillary nerve blocks as a part of multimodal pain management in children undergoing cleft palate repair.4715 Compared with control groups, pediatric patients receiving maxillary nerve blocks had lower postoperative pain scores and reduced opioid consumption.5 The success rates of maxillary pain blocks were not determined in these trials. This is mainly due to the inability to communicate and to perform sensory testing in small children. Although our results indicate that maxillary nerve blocks using a suprazygomatic approach are not always successful, the results from the above-mentioned clinical studies have shown significantly reduced pain levels and opioid consumption in the active treatment groups.5 Consequently, the blocks are a part of a guideline for postoperative pain management.10 However, the number needed to treat to achieve a successful block can range from 1.3 (incomplete block with local anesthetic accumulation in the pterygopalatine fossa on at least one side) to 1.7 (complete block with bilateral spread to the pterygopalatine fossa). It can be speculated whether a local anesthetic injection within the infraorbital fossa (without accumulation in the pterygopalatine fossa) may have a direct effect on other nerve tracts (eg, the alveolar nerves), thereby providing an analgesic effect in patients undergoing palatal repair. So far, however, such mechanisms have not yet been investigated or described. While the use of ultrasound has been advocated by different groups, studies comparing landmark guidance with ultrasound guidance for suprazygomatic maxillary nerve blocks have not yet been published.11 16 Since our current results did not show that ultrasound improves block success, landmark-guided techniques might be a viable option that does not rely on the availability of an ultrasound device. Our findings are particularly relevant for the numerous cleft palate treatments in low-income countries, where ultrasound is not readily accessible.17 Because we did not expect clinically significant differences other than a few seconds for the performance time (defined as the time from skin puncture to local anesthetic injection), we did not include this outcome in our study. Our personal experience with maxillary nerve blocks suggests that ultrasound guidance may cause some delay due to preparation of additional equipment, use of sterile probe covers, and adjustment of ultrasound settings. As our results show a similar local anesthetic distribution, we suggest that the readers choose the most suitable method in their individual clinical setting. Most studies of ultrasound-guided suprazygomatic nerve blocks use an infrazygomatic transducer position. When simulating ultrasound scans using pediatric and adult skull models in a water tank, visualization of the pterygopalatine fossa was almost impossible from this position. With a suprazygomatic transducer position, we were able to visualize the pterygopalatine fossa adequately (figure 1). Critically reviewing previous studies, we note that the deeper part of the infratemporal fossa (superficial to the lateral pterygoid plate and the pterygomaxillary fissure) is often mislabeled as the pterygopalatine fossa in published ultrasound images.11 16 Based on our experience with scanning anatomical models, we perform a suprazygomatic ultrasound scan in our department and used this approach in our study. As an alternative to the suprazygomatic approach, infrazygomatic maxillary nerve blocks have been described using both anterior and posterior needle approaches.18 19 These approaches may potentially improve the accuracy of needle placement. But this could come at the expense of safety, as the needle tip is more exposed to vital structures within the cranium and orbit. The maxillary nerve is located in the pterygopalatine fossa deep to the infratemporal fossa and can be accessed through the pterygomaxillary fissure, a small gap only a few millimeters wide. Hence, a needle inserted into the infratemporal fossa is more likely to hit the lateral pterygoid plate (sphenoid) or the maxilla than enter the pterygopalatine fossa. A local anesthetic injection superficially to the pterygomaxillary fissure is expected to fill the confined space within the infratemporal fossa. As regularly observed with ultrasound guidance, local anesthetic tends to spread superficially where tissue can expand toward the body surface. In our MRI study, local anesthetic spread to the infratemporal fossa was observed in the majority of cases. Yet, deep spread through the pterygomaxillary fissure seemed to be hindered in many block procedures where the injectate did not reach the maxillary nerve. Differences in the needle angles and depths showed some statistically significant differences with p values <0.05. Yet, these differences were relatively small. As the landmark-guided and ultrasound-guided techniques showed similar distribution patterns of the local anesthetic, the differences are unlikely to be clinically relevant. The surgeons who participated in our study had extensive experience with landmark-based suprazygomatic maxillary blocks through their work at Oslo University Hospital and involvement in international surgical missions. Similarly, the anesthesiologists in the study had been using both landmark-based and ultrasound-guided maxillary nerve blocks for several years. The decision to allow either the surgeons or the anesthesiologists to perform the respective guidance technique based on their individual expertise may have introduced some interobserver variability. Our study has several other limitations. First, we do not know if the MRI visualization of injectates reflects true clinical block success. T2-weighted MRI sequences have been used to visualize local anesthetic spread after brachial plexus blocks in previous studies.12 13 However, data on the specificity and sensitivity of the method to predict successful maxillary block in pediatric patients are not available. Furthermore, MRI examination prior to the maxillary nerve block for comparison was not performed in our study. Preprocedural scans were performed on four patients included in a pilot study. Based on these findings, our radiologists were confident that additional examination was not needed for a correct assessment of the postprocedural images. By performing just one MRI scan, we were able to spare the children an unnecessarily long examination under general anesthesia. It must also be taken into account that other fluids, such as blood from an accidental vascular puncture, could be mistakenly identified as local anesthetic injectate. Yet, aspiration of blood was infrequent during our study. We have investigated suprazygomatic maxillary nerve block techniques using a suprazygomatic transducer position for ultrasound-guided blocks. Our results apply to these two specific techniques. Other methods may prove more effective and should be investigated in future trials. We may be criticized for including infants in a trial that could have been conducted on adult patients. Due to their better communication skills and the possibility of sensory testing, an MRI scan might not have been necessary. However, the anatomical structures, such as the shape and thickness of the skull and the maturity of the ligaments and muscles, are fundamentally different in small children.20 Therefore, the spread of local anesthetic and the effect of the block are likely to differ from those in adults. It is important to emphasize that our results apply only to non-syndromic children without craniofacial malformations other than cleft lip and palate. Finally, local anesthetic spread to the pterygopalatine fossa and the maxillary nerve can only be considered a surrogate measure for the expected pain relief, as its relationship to clinical end points remains uncertain. As both ultrasound-guided and landmark-based techniques were performed simultaneously in each patient, our study was not designed to collect clinical data. Therefore, our findings should be confirmed in clinical trials focusing on pain intensity, postoperative agitation, and analgesic consumption as outcome measures. Conclusion In an MRI study of pediatric patients, landmark-guided and ultrasound-guided suprazygomatic maxillary nerve blocks were similarly effective. This suggests that appropriate pain management for cleft palate repair can be achieved without the availability of ultrasound technology. Future studies should focus on improving the success rate of maxillary nerve blocks, both with and without the aid of ultrasound guidance. Supplementary material