Intramasseteric versus pterygomandibular space dexamethasone injection for oedema and trismus after mandibular third molar surgery: a randomised split-mouth clinical trial

Abstract

Background

This study compared intramasseteric versus pterygomandibular space dexamethasone injection for postoperative oedema and trismus after mandibular third molar surgery.

Methods

This prospective, randomised, participant- and outcome assessor-blinded, active-comparator split-mouth clinical trial with sequential interventions was conducted at Afyonkarahisar Health Sciences University between April 28, 2023 and April 27, 2024. Twenty adults aged 18–25 years, classified as American Society of Anesthesiologists Physical Status I and requiring bilateral surgical extraction of comparable mandibular third molars, were enrolled. The two interventions were performed in randomised order at a three-week interval.

Results

All 20 participants completed both interventions. Operative time did not differ between injection sites (18.1 ± 2.8 vs. 18.1 ± 2.9 min; p = 1.000). Change in oedema was lower in the intramasseteric group on days 2 (4 [1–8] vs. 6 [4–12]; p < 0.001; difference between site-specific medians favouring intramasseteric injection, 2.0 mm) and 7 (0.5 [0–1] vs. 1 [0–3]; p = 0.006; difference between medians, 0.5 mm). Trismus was less severe on days 2 (10.4 ± 2.7 vs. 13.1 ± 3.2 mm; p = 0.006; mean difference, 2.7 mm; 95% CI, 0.9 to 4.5) and 7 (0.7 ± 0.8 vs. 1.2 ± 1.2 mm; p = 0.006; mean difference, 0.5 mm; 95% CI, 0.2 to 0.8). Between-group pain, salivary cortisol level, Oral Health Impact Profile-14 score, and analgesic consumption differences were nonsignificant.

Conclusions

Intramasseteric injection was associated with lower oedema and trismus severity. Patient-reported recovery outcomes did not differ significantly between the injection sites. The clinical relevance of these differences remains uncertain.

Trial registration

ClinicalTrials.gov NCT06604286; first submitted 17 September 2024 and first posted 19 September 2024 (retrospectively registered).

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12903-026-09953-5.

Background

Dexamethasone can be administered by systemic routes or direct local infiltration during third molar surgery [1]. Local intraoral protocols have been evaluated for postoperative swelling, pain, and restricted mouth opening in randomised third molar trials [2, 3]. A meta-analysis of four trials found no statistically significant differences in inflammatory outcomes on postoperative day 3 between local and systemic dexamethasone administration [4].

The pterygomandibular space is familiar to oral and maxillofacial surgeons as the route for inferior alveolar nerve block anaesthesia [5]. Direct comparisons between intraoral sites have also been undertaken, including pterygomandibular and sublingual injections [6]. Comparing intramasseteric and pterygomandibular delivery addresses whether the choice of injection site affects recovery when the corticosteroid dose is held constant.

This active-comparator split-mouth clinical trial aimed to evaluate recovery after intramasseteric versus pterygomandibular space dexamethasone injection in bilateral mandibular third molar surgery. We hypothesised that intramasseteric injection would reduce postoperative oedema and trismus. Specific objectives were to compare facial oedema, trismus, pain intensity, oral health-related quality of life, salivary cortisol variation, and analgesic use between the two intraoral injection sites.

Methods

Study design, setting, and samples

To address the research objective, a prospective, randomised, participant- and outcome assessor-blinded, active-comparator split-mouth clinical trial with sequential interventions was conducted at the Department of Oral and Maxillofacial Surgery, Afyonkarahisar Health Sciences University, between April 28, 2023 and April 27, 2024, in accordance with the date range recorded in ClinicalTrials.gov. The study followed the Declaration of Helsinki, was approved by the Afyonkarahisar Health Sciences University Clinical Research Ethics Committee (initial approval: May 13, 2022, no. 2022/267; amendment approval: January 6, 2023, no. 2023/1) before enrolment, and was registered at ClinicalTrials.gov (NCT06604286). The trial was registered retrospectively. All participants provided written informed consent. Reporting followed the CONSORT 2010 statement and the CONSORT extension for within-person randomised trials [7].

The study included adult patients referred for impacted mandibular third molar management during the study period. Eligibility required age 18–25 years, American Society of Anesthesiologists Physical Status I, and asymptomatic, bilaterally symmetric, mesioangular Pell and Gregory Class II-B mandibular third molars requiring surgical extraction. Exclusion criteria were current smoking, pregnancy, lactation, acute pericoronitis or active odontogenic infection, systemic inflammatory disease, regular medication affecting bone metabolism or wound healing, systemic corticosteroid use within 3 months, allergy to local anaesthetics or study drugs, or inability to attend follow-up.

Because directly comparable data for this specific injection-site comparison were limited, sample size was based on a conventional medium standardised mean difference of 0.50 for the primary functional outcome of trismus, expressed relative to a common marginal standard deviation [8]. Equal marginal standard deviations and a within-participant correlation of 0.75 were assumed, corresponding to a paired effect size of d_z = 0.707. An a priori two-sided paired t-test calculation with α = 0.05 and 80% power in G*Power software (v3.0.10; Franz Faul, University of Kiel, Germany) required 18 patients; 20 were enrolled to allow attrition.

Randomisation and blinding

The split-mouth design assigned each participant to both interventions, with intramasseteric injection on one side and pterygomandibular space injection on the contralateral side. Eligible participants were enrolled before allocation. An independent staff member not involved in surgery or outcome assessment generated the computer-based sequence for the initial surgical side and first injection route. The alternate route was used for the contralateral side. Allocation was concealed using sequentially numbered, opaque, sealed envelopes. Each envelope contained both the initial surgical side and the first injection route and was opened before the first operation. The surgeon therefore knew the allocation before surgery, while participants and the outcome assessor remained blinded. A predefined three-week interval separated the surgeries in all participants to allow recovery from the acute postoperative response [9].

All procedures were performed by a single experienced oral and maxillofacial surgeon between 09:00 and 10:00 using a standardised surgical protocol. Local anaesthesia was achieved with 4% articaine containing 1:100,000 epinephrine. After elevation of a full-thickness envelope flap, osteotomy and tooth sectioning were performed under sterile saline irrigation. The sockets were debrided, irrigated with 20 mL sterile saline, and closed with 3 − 0 silk sutures. Postoperative care consisted of amoxicillin/clavulanic acid for 7 days, chlorhexidine/benzydamine mouthwash, and 500-mg paracetamol as rescue analgesia. Local ice application was avoided, and sutures were removed on postoperative day 7.

Variables and data collection methods

The primary predictor was the anatomical site of postoperative dexamethasone administration. Immediately after wound closure, dexamethasone 8 mg (2 mL) was delivered. For the intramasseteric route, the needle was inserted into the masseter muscle body approximately 1 cm superior to the mandibular second molar occlusal plane, with the syringe oriented parallel to that plane and just posterior to the anterior border of the mandibular ramus (Fig. S2A in Additional file 2). For the pterygomandibular route, the syringe was directed from the contralateral canine area toward a point 1 cm superior to the second molar occlusal plane, mirroring an inferior alveolar nerve block trajectory (Fig. S2B in Additional file 2). Aspiration was performed before injection to reduce intravascular administration risk, and the needle was advanced without intentional bone contact.

The primary outcomes were postoperative facial oedema and trismus, recorded by a blinded assessor at baseline and on postoperative days 2 and 7. Oedema was measured with a flexible millimetre ruler using the Gabka-Matsumura linear craniometric method [10], which sums five distances: tragus to lateral canthus, oral commissure, soft-tissue pogonion, and gonion, plus gonion to lateral canthus. Each linear measurement was performed twice by the same blinded assessor, and the mean was analysed; intra-examiner reliability was excellent (ICC = 0.91). Landmarks and pathways are shown in Fig. S1 in Additional file 1. Trismus was evaluated as maximum pain-free interincisal distance using digital callipers.

Secondary outcomes were pain intensity, salivary cortisol, oral health-related quality of life, and analgesic consumption. Pain was measured using a 10-cm visual analogue scale (VAS; 0 = no pain, 10 = worst imaginable pain) at 6, 12, and 24 h and on postoperative days 2 and 7. Unstimulated saliva was collected at baseline and on postoperative days 2 and 7 between 09:00 and 10:00. All saliva samples were collected in sterile tubes and transported immediately to the laboratory. Within 30 min, samples were centrifuged at 3000 rpm for 10 min to remove cellular debris, and the supernatants were stored in sterile cryovials at − 20 °C until analysis. All ELISA assays were performed in a single batch according to the manufacturer’s instructions to minimise inter-assay variability. Cortisol levels were measured using Human Cor ELISA Kits (ELK Biotechnology; sensitivity 1.02 ng/mL). Oral health-related quality of life was evaluated with the validated Turkish OHIP-14 [11]at baseline and on postoperative days 2 and 7, using 0–4 item scoring and a total score range of 0–56. Rescue analgesic use was recorded in patient diaries as paracetamol tablet count. Baseline covariates were age, sex, and body mass index.

Statistical analyses

Statistical analysis was performed using SPSS software (v25.0; IBM Corp., Armonk, NY, USA). Distributional normality was assessed with the Shapiro–Wilk test. Paired-samples t-tests were used for normally distributed paired continuous variables, and Wilcoxon signed-rank tests for nonparametric continuous or ordinal paired variables. Longitudinal within-site changes were assessed with Friedman tests when repeated measures did not satisfy parametric assumptions. Between-site differences in absolute change values were tested using paired t-tests or Wilcoxon signed-rank tests according to distribution. Significance was set at p < 0.05. The split-mouth design controlled for shared participant characteristics; procedural comparability was assessed using operation times. For primary outcomes, between-site estimates were expressed as differences between site-specific medians for nonparametric outcomes and as mean differences for normally distributed paired outcomes. No adjustment for multiple comparisons was applied across secondary outcomes; these results should be interpreted as exploratory.

Results

The CONSORT flow diagram (Fig. 1) illustrates subject screening, allocation, follow-up, and analysis. All 20 participants completed both surgical protocols and follow-up sessions. The sample comprised 10 females (50%) and 10 males (50%) aged 18–25 years. Baseline demographic and clinical characteristics are provided in Supplementary Table S2 in Additional file 4. Between-site effect-size and precision estimates for all primary and secondary outcomes are provided in Supplementary Table S3 in Additional file 4. Repeated facial measurements demonstrated excellent intra-examiner reliability (ICC = 0.91). During the evaluation period, no serious postoperative complications, such as alveolar osteitis, secondary infection, prolonged bleeding, or neurosensory deficits, were observed.

Fig. 1.

CONSORT flow diagram illustrating the clinical sequence of participant screening, randomisation, split-mouth sequence assignment, active interventions during the surgical phase, follow-up intervals, and final statistical analysis

Mean operation times were 18.1 ± 2.8 and 18.1 ± 2.9 min for the intramasseteric injection and pterygomandibular space injection groups, respectively (p = 1.000; Table 1). Median change in oedema from baseline to postoperative day 2 was 4 (1–8) mm in the intramasseteric group and 6 (4–12) mm in the pterygomandibular space group (p < 0.001; difference between site-specific medians favouring intramasseteric injection, 2.0 mm; approximate Wilcoxon r ≥ 0.74; Table 2). On postoperative day 7, the median oedema change from baseline was 0.5 (0–1) mm in the intramasseteric group and 1 (0–3) mm in the pterygomandibular space group (p = 0.006; difference between medians, 0.5 mm; approximate Wilcoxon r = 0.61; Table 2). Within-group comparisons across time points were statistically significant in both groups (p < 0.001).

Table 1.

Comparison of operation times between active sites

Variable	Intramasseteric injection	Pterygomandibular space injection	p-value	Statistical test
Operation time (min)	18.1 ± 2.8	18.1 ± 2.9	1.000	Paired-samples t-test

Values are means ± standard deviations

Table 2.

Changes in postoperative facial oedema from baseline

Time interval	Intramasseteric injection	Pterygomandibular space injection	p-value	Statistical test
Baseline to day 2 (mm)	4 (1–8)	6 (4–12)	< 0.001	Wilcoxon signed-rank test
Baseline to day 7 (mm)	0.5 (0–1)	1 (0–3)	0.006	Wilcoxon signed-rank test
Within-site time effect	p < 0.001	p < 0.001	–	Friedman test

Values are medians (minimum–maximum). Oedema change is expressed in millimetres (mm) and represents the absolute change in the cumulative sum of five linear craniometric measurements. Confidence intervals for differences between site-specific medians were not calculated

Mean reduction in maximum pain-free mouth opening on day 2 was 10.4 ± 2.7 and 13.1 ± 3.2 mm in the intramasseteric and pterygomandibular groups, respectively (p = 0.006; mean between-site difference favouring intramasseteric injection, 2.7 mm; 95% CI, 0.9 to 4.5; Table 3). On day 7, reductions were 0.7 ± 0.8 and 1.2 ± 1.2 mm, respectively (p = 0.006; mean difference, 0.5 mm; 95% CI, 0.2 to 0.8; Table 3). Both sites showed significant temporal variation (p < 0.001). Median VAS scores for intramasseteric versus pterygomandibular injections were 3.5 (0–5) versus 3.5 (0–6) at 6 h, 4.5 (0–7) versus 4.0 (0–7) at 12 h, 5.0 (0–8) versus 5.0 (0–8) at 24 h, 1.5 (0–3) versus 1.0 (0–3) on day 2, and 0 (0–1) versus 0 (0–1) on day 7 (all p > 0.05; Table 4).

Table 3.

Postoperative trismus variations from baseline

Time interval	Intramasseteric injection	Pterygomandibular space injection	p-value	Statistical test
Baseline to day 2 (mm)	10.4 ± 2.7	13.1 ± 3.2	0.006	Paired-samples t-test
Baseline to day 7 (mm)	0.7 ± 0.8	1.2 ± 1.2	0.006	Paired-samples t-test
Within-site time effect	p < 0.001	p < 0.001	–	Friedman test

Values are means ± standard deviations. Positive values indicate a reduction in maximum pain-free interincisal opening from baseline

Table 4.

Postoperative visual analogue scale (VAS) pain scores

Time point	Intramasseteric injection	Pterygomandibular space injection	Between-site p-value	Statistical test
6 h	3.5 (0–5)	3.5 (0–6)	0.718	Wilcoxon signed-rank test
12 h	4.5 (0–7)	4 (0–7)	0.738	Wilcoxon signed-rank test
24 h	5 (0–8)	5 (0–8)	0.989	Wilcoxon signed-rank test
Day 2	1.5 (0–3)	1 (0–3)	0.602	Wilcoxon signed-rank test
Day 7	0 (0–1)	0 (0–1)	1.000	Wilcoxon signed-rank test
Within-site time effect	p < 0.001	p < 0.001	–	Friedman test

Values are medians (minimum–maximum) on a 0–10 VAS

Total OHIP-14 scores, salivary cortisol changes, and cumulative rescue analgesic consumption did not differ significantly between injection sites (all p > 0.05; Table 5). Domain-specific OHIP-14 scores are provided in Supplementary Table S1 in Additional file 3 and showed no significant between-site differences. Sex-specific subgroup analysis was not performed owing to sample size constraints.

Table 5.

Combined secondary postoperative outcomes

Secondary outcome	Intramasseteric injection	Pterygomandibular space injection	Between-site p-value	Statistical test
OHIP-14 total, baseline	6 (0–30)	6 (0–32)	1.000	Wilcoxon signed-rank test
OHIP-14 total, day 2	11 (0–39)	14 (0–43)	0.414	Wilcoxon signed-rank test
OHIP-14 total, day 7	6 (0–34)	6 (0–36)	0.925	Wilcoxon signed-rank test
Salivary cortisol change, baseline to day 2 (ng/mL)	7.5 ± 9.6	5.6 ± 9.4	0.351	Paired-samples t-test
Salivary cortisol change, baseline to day 7 (ng/mL)	3.6 ± 7.1	1.5 ± 10.4	0.552	Paired-samples t-test
Total paracetamol tablets	9 (6–11)	9 (6–11)	0.624	Wilcoxon signed-rank test

Values are medians (minimum–maximum) for OHIP-14 total scores and rescue analgesic use, and means ± standard deviations for salivary cortisol concentration changes. OHIP-14 total scores were calculated using 0–4 item scoring (total range, 0–56). Cortisol concentrations are reported in ng/mL and were measured using enzyme-linked immunosorbent assays. Rescue analgesic use represents cumulative 500-mg paracetamol tablet consumption. OHIP-14 = Oral Health Impact Profile-14. Detailed OHIP-14 domain scores are provided in Supplementary Table S1 in Additional file 3

Discussion

This active-comparator split-mouth clinical trial evaluated postoperative recovery after intramasseteric or pterygomandibular space dexamethasone injection following mandibular third molar surgery. Results supported the hypothesis that intramasseteric injection would reduce objective postoperative oedema and trismus on days 2 and 7. However, patient-reported pain intensity, OHIP-14 scores, salivary cortisol levels, and rescue analgesic use did not differ significantly between injection-site groups.

The masseter muscle lies immediately lateral to the mandibular ramus and contributes to postoperative inflammatory response after mandibular third molar removal. Dexamethasone is effective for acute postoperative swelling [12, 13]. Dereci et al. reported less day-2 swelling after postoperative intramasseteric dexamethasone than after saline injection in a 20-participant split-mouth trial [14]. Rocha-Neto et al. reported less oedema and trismus after immediate postoperative intramasseteric injection of 8 mg dexamethasone than after surgery without corticosteroid administration, although pain did not differ significantly [2]. In a 210-participant randomised trial, Gupta et al. found less swelling and mouth-opening restriction on days 1 and 3 with intramasseteric than intradeltoid administration of 8 mg dexamethasone given 2 h before surgery [15]. Our comparison extends these observations to two intraoral sites receiving the same postoperative dose, with patient-reported recovery assessed alongside oedema and trismus.

The patterns of trismus directly mirrored those of oedema, with the intramasseteric group showing significantly lower reduction in maximum mouth opening at both follow-up points. Greater day-2 mouth-opening limitation after pterygomandibular than sublingual dexamethasone injection was also reported in a split-mouth trial [6]. No serious postoperative complications were observed in the present study, although the small sample limits assessment of injection-related harms, including intramuscular haematoma.

An important interpretive challenge is the discrepancy between objective and patient-reported outcomes. Although swelling and trismus were statistically reduced after intramasseteric administration, perceived pain, quality-of-life impacts, and rescue analgesic requirements did not differ significantly. Day-2 mouth-opening reductions averaged 10.4 and 13.1 mm, with a between-site difference of 2.7 mm that declined to 0.5 mm on day 7. The oedema contrast concerned the sum of five facial distances, with differences between site-specific medians of 2.0 mm on day 2 and 0.5 mm on day 7. A patient-important threshold was not prespecified for either measure, leaving the clinical relevance of these differences uncertain.

OHIP-14 has demonstrated responsiveness to changes during recovery from third molar surgery [16, 17]. Its general oral-health domains may not capture small differences in acute symptoms between two active corticosteroid regimens. A procedure-specific instrument such as the Postoperative Symptom Severity scale could complement OHIP-14 by assessing symptoms and disruption of daily activities after third molar surgery [18].

The main limitation was the lack of inactive placebo/saline or systemic comparator arms. This active-comparator trial focused on comparing local intraoral methods when dexamethasone administration is planned. With only 20 participants, the study may be underpowered for smaller trismus effects than assumed and for subtle secondary behavioural or endocrine differences, precluding sex- or body mass index-related multivariable analyses. The single-centre setting and performance of all procedures by a single experienced surgeon improved procedural consistency but may limit external generalisability to operators with different experience levels or institutional protocols. The surgeon knew the treatment allocation, leaving potential for performance bias despite procedural standardisation. The narrow age range and inclusion of only ASA I young adults also limit extrapolation to older patients or individuals with systemic comorbidities undergoing third molar surgery. Restriction to mesioangular Class II-B impactions limits applicability to other impaction patterns; whether greater inflammatory severity would alter the relationship between objective and patient-reported recovery remains uncertain. Although the three-week interval was intended to reduce carry-over, complete inflammatory or endocrine recovery before the second session was not biochemically confirmed. Dental anxiety, sleep quality, and session-specific stressors were not measured; the experience of the first operation may have influenced patient-reported recovery and cortisol measurements during the second session through altered expectations or anxiety. Both teeth were classified as Pell and Gregory Class II-B, and mean operation times were similar, although these measures may not capture all differences in surgical difficulty. Period and treatment-sequence effects were not formally analysed, leaving uncertainty about their contribution to the observed differences. Retrospective registration limits independent verification of the prespecified outcomes and analyses.

Although repeated linear facial measurements showed excellent intra-examiner reliability (ICC = 0.91), the Gabka-Matsumura method provides a linear proxy for swelling; three-dimensional stereophotogrammetry permits volumetric assessment [19]. Future multiarm randomised trials should include saline/placebo and systemic comparator arms, digital facial scanning, and predefined minimal clinically important differences for facial expansion and mandibular restriction to clarify whether the observed between-route differences offer patient-relevant clinical value.

Conclusions

Intramasseteric dexamethasone injection was associated with lower objective postoperative oedema and trismus than pterygomandibular space injection after mandibular third molar surgery. The measured differences were not accompanied by statistically significant differences in patient-reported pain, oral health-related quality of life, salivary cortisol response, or analgesic use.

Abbreviations

ASA

American Society of Anesthesiologists

BMI

Body mass index

CI

Confidence interval

CONSORT

Consolidated Standards of Reporting Trials

ELISA

Enzyme-linked immunosorbent assay

ICC

Intraclass correlation coefficient

OHIP-14

Oral Health Impact Profile-14

VAS

Visual analogue scale

Authors’ contributions

M.Y.P. contributed to conceptualization, study design, methodology, investigation, data curation, formal analysis, drafting of the original manuscript, and critical revision of the manuscript. O.T. contributed to methodology, investigation, data curation, formal analysis, and critical revision of the manuscript. Both authors read and approved the final manuscript.

Funding

This work was supported by the Afyonkarahisar Health Sciences University Scientific Research Projects Unit (project no. 23.DUS.002). The funder had no role in study design, data collection, analysis, interpretation, manuscript preparation, or the decision to submit the article for publication.

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to participant privacy and ethical restrictions but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki and was approved by the Afyonkarahisar Health Sciences University Clinical Research Ethics Committee before participant enrolment (initial approval: May 13, 2022; approval no. 2022/267; amendment/project conversion approval: January 6, 2023; meeting no. 2023/1). Written informed consent was obtained from all participants.

Consent for publication

Consent for publication was obtained from all participants whose clinical images are included. The facial reference image was masked/de-identified; no intraoral procedural photographs are included in the manuscript.

Competing interests

The authors declare no competing interests.