Diagnostic re-evaluation of non-odontogenic orofacial pain after treatment failure: a retrospective case series

2.1 Study design and population

This was a retrospective case series. Among the 1824 new patients who presented to the Department of Dentistry and Oral and Maxillofacial Surgery of a regional general hospital between April 2023 and April 2025, 54 had a chief complaint of orofacial pain and were referred from community dental clinics for further evaluation and management due to suspected non-odontogenic orofacial pain.

2.2 Standardized provocation tests

Percussion testing involved vertical and lateral taps with a mouth mirror handle, occlusal testing involved a cotton roll and natural occlusion, and thermal testing used ethyl chloride spray and warm water (40 °C). Failure to reproduce the patient’s usual pain with any provocation was defined as “provocation negative”.

Pain intensity was assessed using an 11-point NRS (0–10), with clinically meaningful change defined as ≥2-point improvement.

2.3 Neuromuscular and chairside sensory examination

All patients underwent systematic musculoskeletal and neurosensory evaluations as part of the standardized clinical re-evaluation protocol. Musculoskeletal examination included the following: (1) Bilateral palpation of the masticatory muscles (masseter and temporalis) to assess tenderness, muscle stiffness, and trigger points; (2) Bilateral palpation of the temporomandibular joints (TMJs), including the lateral pole and posterior attachment areas; (3) Trigger point provocation testing using the application of standardized moderate digital pressure for several seconds to determine whether the patient’s familiar pain could be reproduced; (4) Assessment of side-to-side differences in muscle tenderness and TMJ findings, involving neurosensory examination conducted using chairside sensory testing, rather than formal instrument-based quantitative sensory testing (QST).

Specifically, neurosensory examination comprised the following assessments: (i) Light-touch detection involving qualitative assessment using a cotton swab applied bilaterally to the region corresponding to the trigeminal nerve distributions; (ii) Touch-pressure detection, evaluated using Semmes–Weinstein monofilaments when somatosensory abnormality was suspected; (iii) Mechanical pain sensitivity, examined using a sterile dental explorer (pinprick test) with gentle stimulation across the affected and contralateral sides. All sensory assessments were performed using side-to-side comparison to identify asymmetry, including hypoesthesia (reduced sensation), hyperesthesia (increased sensation), or allodynia (pain from non-noxious stimuli). These chairside assessments were conducted to screen clinically relevant somatosensory abnormalities, as well as to guide diagnostic reclassification in conjunction with clinical history, provocation test findings, and imaging results. Standardized computerized QST equipment was not available in the clinical setting. Thermal provocation testing (ethyl chloride spray for cold, warm water at approximately 40 °C) was conducted as part of dental provocation tests (described in Section 2.2), rather than as systematic thermal threshold assessment.

2.4 Diagnostic workflow and case sorting

Based on a retrospective review of referred cases with unresolved orofacial pain, we organized the clinical decision points and re-evaluation steps into a conceptual diagnostic workflow. This framework was not applied as a predefined protocol at the time of initial diagnosis, but represents a post-hoc synthesis of the clinical reasoning processes used across cases. The resulting diagnostic and re-evaluation framework is illustrated in Fig. 1.

Fig. 1.Algorithm for diagnosis and re-evaluation of non-odontogenic orofacial pain. Initial assessment includes systematic history (pain characteristics, duration, triggers, associated symptoms), clinical examination (intraoral exam, imaging, provocation tests), and musculoskeletal/neurological evaluation. Red flags: (1) persistent pain + objective negatives; (2) ipsilateral autonomic signs (nasal congestion, conjunctival injection, tearing); (3) nonresponse to usual dental treatment. When present, pause irreversible procedures and proceed with differential diagnosis per ICOP/ICHD-3/DC/TMD; refer classic cases (paroxysmal electric pain; TACs signs) to appropriate specialists. If red flags are absent and clear odontogenic pathology is identified, provide standard dental care, reassess NRS every 2–4 weeks, and return to re-evaluation if improvement is insufficient. ICOP: International Classification of Orofacial Pain; ICHD-3: International Classification of Headache Disorders, 3rd edition; DC/TMD: Diagnostic Criteria for Temporomandibular Disorders; NSAIDs: nonsteroidal anti-inflammatory drugs; TACs: trigeminal autonomic cephalalgias; PIDP: persistent idiopathic dentoalveolar pain; PDAP: persistent dentoalveolar pain disorder; NRS: numerical rating scale; RCT: Root Canal Treatment; TN: Trigeminal Neuralgia; TCA: Tricyclic Antidepressant; PT: Physical Therapy.

Using this workflow, cases were triaged and re-evaluated as follows: Among the 54 referred cases, 50 were confirmed to have non-odontogenic pain, while 4 showed odontogenic pathology and were returned to the referring clinics. Of the 50 non-odontogenic cases, 41 showed a rapid response to first-line pharmacotherapy or management protocols specific to their diagnoses (e.g., carbamazepine for typical trigeminal neuralgia, splint therapy for temporomandibular disorders) in our department. In three patients, primary headache was strongly suspected at the initial visit, and they were directly referred to the neurology department. Six cases showed difficulty with definitive diagnosis or symptom control and were designated as the analytical cohort. Among these, one case (Case 4) was referred to the neurology department after ineffective initial treatment and was subsequently diagnosed with trigeminal autonomic cephalalgias (TACs). In total, four cases were referred to the neurology department.

2.5 Diagnostic procedure

Final diagnoses were assigned independently by two oral and maxillofacial surgeons (board-certified specialists with 30 years of experience and board-certified dentists with 6 years of experience) based on the ICOP, ICHD-3, and Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) operational criteria. For PIDP, the ICOP criteria were applied as follows: (1) pain persisting or recurring for >3 months, (2) pain located in the dentoalveolar region for >2 h per day, (3) no clear odontogenic or other pathology, and (4) lack of response to usual dental treatments. Any discrepancies were resolved by consensus.

2.6 Data collection

We extracted the following data from the medical records: demographics, medical history, clinical findings (site/quality of pain and associated symptoms), imaging findings, neurological/musculoskeletal findings, treatments, and treatment responses. Follow-up data were collected until August 2025, with some patients remaining under observation at that time. Medications were recorded using generic names. Regarding kampo co-therapy, we recorded its presence/absence, formula names, and treatment response. In some cases, kampo served as an adjunctive background therapy and was excluded from the primary efficacy analyses. Tooth numbering was performed using the FDI two-digit tooth numbering system (FDI system). Diagnostic delay (months) was defined as the period from symptom onset to final diagnosis (or provisional diagnosis) at our department. Missing items in referral letters/records were labeled “NA”. Table 1 summarizes the evaluation items (variable definitions and denominators in the footnotes).

To facilitate systematic diagnostic re-evaluation and enable transparent cross-case comparison, a structured data extraction framework (Table 1) was applied. Table 1 summarizes the core clinical variables and operational definitions used in this study, organized into nine categories: background information, pain-related information, imaging findings, provocation tests, treatment history, pain intensity, neuromuscular findings, autonomic symptoms, and diagnostic delay. This standardized approach ensured consistent evaluation across all cases, and provided the basis for identifying recurring diagnostic pitfalls.

2.7 Identification of misdiagnosis factors

Based on a qualitative review of the six cases, factors contributing to misdiagnosis were identified and categorized into four domains through an author-derived clinical synthesis. This process was informed by established concepts in the diagnostic error literature, including cognitive biases (e.g., anchoring and premature closure), incomplete clinical assessment, and contextual factors [7, 11, 17, 18]. This framework was exploratory in nature and was not intended to represent a previously validated classification system.

2.8 Statistical analysis

Only descriptive statistics were used. All statistical analyses were performed using spreadsheet software (Microsoft Excel, Microsoft 365; Microsoft Corp., Redmond, WA, USA).

2.9 Case descriptions

An overview of the clinical courses and final diagnoses for the six cases is provided in Table 2.

4.1 Misdiagnosis factors and case-derived insights

Using this exploratory framework, four recurring domains of misdiagnosis were identified across the analyzed cases (Table 3). In particular, the misdiagnosis domains corresponded to: (1) overreliance on imaging and prior medication history, (2) overlooking autonomic signs, (3) insufficient musculoskeletal or neurological examination, and (4) premature attribution to a single etiology [7, 8, 9, 10, 11]. The three proposed red flags (stop points) applied to at least one item in every case, which suggested practical triggers for re-evaluation. However, given the small sample size (n = 6), their diagnostic performance requires prospective validation.

4.2 Cognitive biases potentially involved

Anchoring (fixation on an initial “odontogenic” hypothesis with underweighting of conflicting negatives), confirmation bias (selective search for supportive data), availability heuristic (overestimation of common odontogenic diseases), and overreliance on imaging were all suggested in our cases [7, 11, 17, 18]. Imaging should remain adjunctive and integrated with reproducible neurological and musculoskeletal findings.

4.3 Case-specific lessons

• Case 1: Persistent spontaneous pain with negative imaging/provocation following irreversible procedures should prompt re-evaluation of PIDP/PDAP [15, 19, 20].

• Case 2: Temporal linkage to dental extraction, unilateral lingual localization, and responsiveness to neuropathic agents all favored a diagnosis of traumatic lingual nerve injury [21, 22, 23] over chemotherapy-associated neuropathy [24]; however, the latter could not be fully excluded. Although the patient had a history of chemotherapy, their clinical presentation was more consistent with a localized traumatic trigeminal neuropathy than with chemotherapy-related neuropathy. The sensory disturbance was strictly confined to the lingual nerve distribution, unilateral, and demonstrated clear side-to-side differences. This sharply localized, single-branch trigeminal pattern is less typical for chemotherapy-related neuropathy, which more commonly presents as diffuse or multifocal sensory symptoms, rather than a focal cranial nerve distribution. Furthermore, symptom onset and persistence were temporally associated with the preceding dental extraction, further supporting a diagnosis of localized traumatic nerve injury, rather than systemic chemotherapy effects [24].

• Case 3: The patient’s initial diagnostic trajectory was strongly influenced by clinical context. She had a history of oral bisphosphonate use and underwent tooth extraction in February 2023, followed by additional curettage in May 2023. Pain remained consistently localized to the extraction site, and in the early disease stage, the patient described the pain as persistent, rather than paroxysmal. Given the presence of bisphosphonate exposure and ongoing post-extraction pain following repeated invasive procedures, MRONJ was initially considered the most plausible diagnosis, which was deemed clinically reasonable based on existing guidelines and risk stratification [25]. The predominance of invasive dental procedures and localized nociceptive input likely masked the underlying neural mechanism, as well as the absence of a clearly reported electric shock-like pain in the early phase, reduced the suspicion of trigeminal neuralgia. At the initial stage, pain assessment focused primarily on pain location and persistence, while key phenotypic features—such as attack duration, pain triggered by light mechanical stimuli, presence of brief paroxysms and trigger zones—were not systematically explored or documented. This limited the early recognition of any neural pain pattern. Upon systematic re-evaluation, critical features that had not been previously emphasized emerged, including brief paroxysmal attacks lasting seconds, electric shock-like quality, mechanical triggers during oral contact, and well-defined trigger zones. These findings fulfilled the diagnostic criteria for classical trigeminal neuralgia according to the ICHD-3 [3, 26], and were further supported by the rapid and sustained response to carbamazepine [27]. This case further illustrates a diagnostic pitfall in which anchoring to a high-risk medication history and local dental pathology may delay the consideration of trigeminal neuralgia, particularly when early pain descriptors are nonspecific and confined to the site of dental intervention.

• Case 4: Autonomic signs (nasal congestion and conjunctival injection) prompted early recognition and referral for TACs [2, 3, 16, 28].

• Case 5: Diagnosis and dominant mechanisms changed over time, which argued against premature closure and explicit on-hold diagnosis with iterative reassessment [7, 11, 17, 18]. Finally, comorbid neuropathic and myofascial pain was suspected [12, 29, 30].

• Case 6: Vascular contact on MRI alone should not determine trigeminal neuralgia [3, 26, 27]. Lack of classic paroxysms/trigger points and response to NSAIDs/muscle relaxants/rehabilitation are suggestive of myofascial pain [12, 29, 30]. However, a neuropathic contribution could not be fully excluded, given the long pain history and transient residual pressure-evoked tenderness despite resolution of spontaneous pain. This case illustrates the diagnostic complexity when neuropathic and musculoskeletal pain mechanisms may coexist [29].

4.4 Clinical significance of case 5: dynamic diagnostic reconstruction

Over a 6-year course, the diagnosis evolved from suspected PDAP to overt myofascial features and finally to comorbid neuropathic and myofascial pain. Treatment responses (tricyclics → carbamazepine → celecoxib) suggest shifting dominant mechanisms [13, 29]. Declaring “diagnosis pending”, with periodic re-evaluation and pharmacologic provocation as quasi-diagnostic probes, may improve accuracy [11, 17, 18]. For chronic orofacial pain, it is more realistic to seek functional improvement and acceptable pain, rather than complete remission [13, 30].

4.5 Pharmacotherapy: role and limits

Medication response informs, but should not dictate the diagnosis. For example, the considerable effect of carbamazepine in Case 3 supports (but does not solely establish) trigeminal neuralgia in conjunction with paroxysms and trigger points [3, 26, 27]. Pharmacological management should be guided by the final clinical diagnosis established using validated diagnostic criteria, rather than by the pain quality alone [2, 3, 12]. However, qualitative pain characteristics may provide supportive information during diagnostic reasoning [13]. For example, paroxysmal electric shock-like pain may raise the suspicion of trigeminal neuralgia [26, 27], while persistent burning or dysesthetic pain may indicate a neuropathic pain component [2, 13, 15]. Similarly, localized muscle tenderness or chewing-related pain may indicate a possible myofascial contribution [12, 30]. Importantly, these features were used only as adjunctive clues within a comprehensive diagnostic framework, not as standalone determinants of treatment selection.

Once a diagnosis has been established using validated criteria (ICOP, ICHD-3, DC/TMD), an evidence-based pharmacotherapy strategy can be selected accordingly [2, 3, 12]. For example, trigeminal neuralgia diagnosed per the ICHD-3 criteria is treated with carbamazepine as a first-line therapy [3, 26, 27]. Persistent idiopathic dentoalveolar pain meeting the ICOP criteria may be managed with gabapentinoids or tricyclic antidepressants based on evidence for neuropathic pain conditions [2, 13, 15, 19, 20]. Myofascial pain diagnosed per the DC/TMD criteria is treated with NSAIDs, muscle relaxants, and physical therapy [12, 29, 30]. Treatment response should inform, but not replace, systematic diagnostic evaluation, as multiple pain mechanisms may coexist (as demonstrated in Case 5) [29].

Sequential or combination pharmacotherapy may be necessary for cases of complex chronic pain [13]. However, in our practice, each medication trial was predefined with a fixed evaluation window (typically 2–4 weeks) and an a priori response threshold (≥2-point NRS improvement), to facilitate structured reassessment. Pharmacotherapy alone rarely yields complete remission, while rehabilitation and self-care are often essential [29, 30].

4.6 Implementing ”stop points” and a re-evaluation algorithm

The proposed “red flags” for diagnostic re-evaluation were primarily derived from the authors’ collective clinical experience in the management of treatment-refractory orofacial pain, and were informed by established textbooks and prior literature, rather than validated diagnostic criteria, and are illustrated in Fig. 1. Our two-stage exclusion (triage at referral and confirmation after the first visit) and independent dual-rater assignment with consensus may facilitate operationalization. Based on the ICOP/ICHD-3/DC/TMD criteria [2, 3, 12] and existing literature [7, 8, 9, 15], we propose three red flags:

1. Persistent pain objective negatives: No clear odontogenic pathology on imaging, no acute inflammatory signs, or negative provocation tests.

2. Atypical associated signs: ipsilateral nasal congestion, conjunctival injection, and tearing (autonomic signs).

3. Incongruent treatment response—insufficient response to usual dental management and/or atypical response patterns suggesting an alternative or mixed pain mechanism.

Conceptually, these red flags integrate the established diagnostic principles (ICOP for persistent dentoalveolar pain, ICHD-3 for trigeminal autonomic cephalalgias, and prior reports of treatment-refractory non-odontogenic pain) with our clinical observations [2, 3, 7, 8, 9, 15]. While the individual components have been well-described in the literature, their combination as a practical framework for triggering diagnostic re-evaluation represents our clinical synthesis and has not been prospectively validated.

In our six cases, there was at least one red flag per case. These indicators are hypothesis-generating and intended to function as clinical triggers prompting suspension of further irreversible dental procedures and the initiation of systematic diagnostic re-evaluation, rather than as standalone diagnostic tools. The presence of these factors should, therefore, prompt reassessment according to the ICOP/ICHD-3/DC/TMD criteria, including history, provocation testing, musculoskeletal and neurological examination, and evaluation of headache-associated signs [2, 3, 12]. Conversely, diagnoses of classic trigeminal neuralgia or TACs should prompt early specialty referral [3, 26, 27, 28].

4.7 Educational and clinical implications

Misdiagnosis and prolonged diagnostic delay in non-odontogenic orofacial pain impose substantial burdens on patients beyond physical symptoms. Qualitative studies have shown that patients with persistent orofacial pain frequently experience frustration, anxiety, erosion of trust in healthcare providers, and a sense of being dismissed or disbelieved when pain persists despite repeated interventions [31]. In our case series, the median diagnostic delay of 17 months and the repeated performance of irreversible dental procedures exemplify this problem. Diagnostic errors in medicine are often driven by cognitive biases, including anchoring and premature closure, which may perpetuate incorrect diagnostic trajectories once an initial diagnostic label has been applied [32].

The primary clinical implication of our study is the potential to reduce “diagnostic momentum”, where a patient is passed from one clinician to another with an incorrect label, leading to a cascade of unnecessary procedures. Our findings indicate that the proposed “stop-and-re-evaluate” protocol is not merely a specialist tool, but a fundamental safeguard in general dental practice. Implementing this pause when “red flags” are present could significantly minimize the physical burden of irreversible treatments (e.g., pulpectomies and extractions) and the psychological distress associated with unresolved chronic pain. Furthermore, this study highlights the necessity of interdisciplinary collaboration. As observed in Cases 4 (TACs) and 6 (myofascial pain), the differential diagnosis commonly spans across the neurology and musculoskeletal domains. Therefore, the dental education curricula should place greater emphasis on the “medical” model of orofacial pain assessment, moving beyond the traditional “surgical/restorative” model. Establishing clear referral pathways between general dentists, oral surgeons, and neurologists is essential for the early management of conditions like TACs and trigeminal neuralgia.

4.8 Limitations and future directions

This study has several important limitations. First, as a single-center retrospective case series, the present study population was highly selected, consisting of referred patients following diagnostic difficulty or treatment failure. This referral pattern likely overrepresents complex or treatment-refractory cases with prolonged diagnostic delay, while underrepresenting patients who were successfully diagnosed and managed in general dental practice. Consequently, the distribution of diagnoses and diagnostic delay observed in this series is likely inapplicable to the broader dental population. Second, the retrospective design potentially introduced information bias, as clinical data were derived from medical records, referral letters, and patient recall, which varied in completeness and accuracy. This may have affected the assessment of symptom onset, prior treatments, and factors contributing to misdiagnosis. Third, comprehensive standardized QST using computerized equipment was not available in our clinical setting. Consequently, somatosensory evaluation relied on standardized chairside sensory testing, which may have limited full characterization of sensory profiles in patients with suspected neuropathic pain. Fourth, the proposed “red flags” represent hypothesis-generating indicators derived from the integration of clinical experience with established diagnostic concepts, rather than validated diagnostic criteria. Their diagnostic performance characteristics, including sensitivity, specificity, and predictive values, remain unknown and require prospective validation. Fifth, the small sample size (n = 6) and the heterogeneity of final diagnoses (persistent idiopathic dentoalveolar pain, traumatic neuropathy, trigeminal neuralgia, TACs, and myofascial pain) limited our ability to draw condition-specific conclusions. Furthermore, treatment approaches and follow-up durations varied across cases, precluding systematic evaluation of long-term outcomes. Accordingly, the findings of this case series should be interpreted as only hypothesis-generating. As such, future multicenter prospective studies with larger, unselected patient populations and standardized diagnostic protocols are warranted to validate the proposed clinical framework and red flags, to establish their diagnostic performance characteristics, and to assess their clinical utility in routine dental practice.